Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “AUTOS”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 505 records · Page 28

Cosmology from Planck CMB lensing and DESI DR1 quasar tomography

We present a measurement of the amplitude of matter fluctuations over the redshift range 0.8 ≤ z ≤ 3.5 from the cross correlation of over 1.2 million spectroscopic quasars selected by the Dark Energy Spectroscopic Instrument (DESI) across 7,200 deg 2 (∼ 170 quasars/deg 2 ) and Planck PR4 (NPIPE) cosmic microwave background (CMB) lensing maps. We perform a tomographic measurement in three bins centered at effective redshiftsz=1.44, 2.27 and 2.75, which have ample overlap with the CMB lensing kernel. From a joint fit using the angular clustering of all three redshift bins (auto and cross-spectra), and including an Q m prior from DESI DR1 baryon acoustic oscillations to break the $Ω_{m}-σ_{8}$ degeneracy, we constrain the amplitude of matter fluctuations in the matter-dominated regime to be $σ_{8}=0.929^{+0.059}_{-0.074}$ and $S_{8}≡σ_{8}(Ω_m/0.3)^{0.5} = 0.922^{+0.059}_{-0.073}$. We provide a growth of structure measurement with the largest spectroscopic quasar sample to date at high redshift, which is ∼ 1.5σ higher than predictions from ΛCDM fits to measurements of the primary CMB from Planck PR4. The cross-correlation between PR4 lensing maps and DESI DR1 quasars is detected with a signal-to-noise ratio of 21.7 and the quasar auto-correlation at 27.2 for the joint analysis of all redshift bins. We combine our measurement with the CMB lensing auto-spectrum from the ground-based Atacama Cosmology Telescope (ACT DR6) and Planck PR4 to perform a sound-horizon-free measurement of the Hubble constant, yielding $H_{0}=69.1^{+2.2}_{-2.6} {\rm \ km\ s}^{-1}{\rm Mpc}^{-1}$ through its sensitivity to the matter-radiation equality scale.

baryon acoustic oscillations↗

Mitigating foreground bias to the CMB lensing power spectrum for a CMB-HD survey

A promising way to measure the distribution of matter on small scales (k~10 h Mpc –1 ) is to use gravitational lensing of the cosmic microwave background (CMB). CMB-HD, a proposed high-resolution, low-noise millimeter survey over half the sky, can measure the CMB lensing auto spectrum on such small scales enabling measurements that can distinguish between a cold dark matter (CDM) model and alternative models designed to solve problems with CDM on small scales. However, extragalactic foregrounds can bias the CMB lensing auto spectrum if left untreated. Here, we present a foreground mitigation strategy that provides a path to reduce the bias from two of the most dominant foregrounds, the thermal Sunyaev-Zel’dovich effect and the cosmic infrared background. Given the level of realism included in our analysis, we find that the thermal Sunyaev-Zel’dovich alone and the cosmic infrared background alone bias the lensing auto spectrum by 0.6σ and 1.1σ, respectively, in the lensing multipole range of Lϵ[5000,20000] for a CMB-HD survey; combined these foregrounds yield a bias of only 1.3σ. Including these foregrounds, we also find that a CMB-HD survey can distinguish between a CDM model and a 10 –22 eV fuzzy dark matter model at the 5σ level. These results provide an important step in demonstrating that foreground contamination can be sufficiently reduced to enable a robust measurement of the small-scale matter power spectrum with CMB-HD.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Mixed gain detector configurations for time-resolved X-ray solution scattering

X-ray detection at X-ray free-electron lasers is challenging in part due to the XFEL's extremely short and intense X-ray pulses. Experimental measurements are further complicated by the large fluctuations inherent to the self-amplified spontaneous emission process producing the X-rays. At the Linac Coherent Light Source the ePix10ka2M detector offers multiple gain modes, and auto-ranging between these, to increase the dynamic range while retaining low noise. For diffuse scattering techniques, such as time-resolved X-ray solution scattering, where the shape of the scattering pattern largely does not change between exposures, a fixed mix of different gain modes offers many of the same advantages as auto-ranging. We find that configuring individual ASICs in separate gain modes does not impact the intensity linearity of the gain response and has a limited effect on the effective dynamic range in regions with different gain mode settings while avoiding the complexities of auto-ranging. Small (<5%) non-linear gain contributions arise when pixels on the same ASIC are configured in different gain modes. We present a configuration scheme that is designed to select the optimal mixed gain configuration to minimize effects of saturation in the high-/medium-gain region, while maximizing the number of pixels with higher gain to improve the signal-to-noise ratio.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Automatic Generation of High-Performance Convolution Kernels on ARM CPUs for Deep Learning

In this work, we present FastConv, a template-based code auto-generation open source library that can automatically generate high-performance deep learning convolution kernels of arbitrary matrices/tensors shapes. FastConv is based on the Winograd algorithm, which is reportedly the highest performing algorithm for the time-consuming convolution layers of convolutional neural networks. ARM CPUs cover a wide range designs and specifications, from embedded devices to HPC-grade CPUs. The leads to the dilemma of how to consistently optimize Winograd-based convolution solvers for convolution layers of different shapes. FastConv addresses this problem by using templates to auto-generate multiple shapes of tuned kernels variants suitable for skinny tall matrices. As a performance portable library, FastConv transparently searches for the best combination of kernel shapes, cache tiles, scheduling of loop orders, packing strategies, access patterns, and online/offline computations. Auto-tuning is used to search the parameter configuration space for the best performance for a given target architecture and problem size. The experiments with layer-wise evaluation on the VGG--16 model confirms a 1.25x performance gains is got by tuning the Winograd library. Integrated comparison results shows 1.02x to 1.40x, 1.14x to 2.17x, and 1.22x and 2.48x speedup is achieved over NNPACK, Arm NN, and FeatherCNN on the Kunpeng 920 beside few cases. Furthermore, problem size performance portability experiments with various convolution shapes shows that FastConv achieves 1.2x to 1.7x speedup and 2x to 22x speedup over NNPACK and ARM NN inference engine using Winograd on Kunpeng 920 . CPU performance portability evaluation on the VGG--16 show an average speedup over NNPACK of 1.42x, 1.21x, 1.26x, 1.37x, 2.26x, and 11.02x is observed on Kunpeng 920, Snapdragon 835, 855, 888, Apple M1, and AWS Graviton2, respectively.

97 MATHEMATICS AND COMPUTING↗

A cross-correlation study of the Fermi -LAT γ-ray diffuse extragalactic signal

In this work, starting from 21 months of data from the Fermi Large Area Telescope (LAT), we derive maps of the residual isotropic γ-ray emission, a relevant fraction of which is expected to be contributed by the extragalactic diffuse γ-ray background (EGB). We search for the auto-correlation signals in the above γ-ray maps and for the cross-correlation signal with the angular distribution of different classes of objects that trace the large-scale structure of the Universe. We compute the angular two-point auto-correlation function of the residual Fermi -LAT maps at energies E > 1 GeV, E > 3 GeV and E > 30 GeV well above the Galactic plane and find no significant correlation signal. This is, indeed, what is expected if the EGB were contributed by BL Lacertae (BLLacs), Flat Spectrum Radio Quasars (FSRQs) or star-forming galaxies, since, in this case, the predicted signal is very weak. Then, we search for the Integrated Sachs–Wolfe (ISW) signature by cross-correlating the Fermi-LAT maps with the 7-year Wilkinson Microwave Anisotropy Probe ( WMAP7 ) cosmic microwave background map. We find a cross-correlation consistent with zero, even though the expected signal is larger than that of the EGB auto-correlation. Lastly, in an attempt to constrain the nature of the γ-ray background, we cross-correlate the Fermi -LAT maps with the angular distributions of objects that may contribute to the EGB: quasi-stellar objects (QSOs) in the Sloan Digital Sky Survey Data Release 6 (SDSS-DR6) catalogue, NRAO VLA Sky Survey (NVSS) galaxies, Two Micron All Sky Survey (2MASS) galaxies and Luminous Red Galaxies (LRGs) in the SDSS catalogue. The cross-correlation is always consistent with zero, in agreement with theoretical expectations, but we find (with low statistical significance) some interesting features that may indicate that some specific classes of objects contribute to the EGB. A χ 2 analysis confirms that the correlation properties of the 21-month data do not provide strong constraints of the EGB origin. However, the results suggest that the situation will significantly improve with the 5- and 10-yr Fermi -LAT data. In future, the EGB analysis will then allow placing significant constraints on the nature of the EGB and might provide, in addition, a detection of the ISW signal.

79 ASTRONOMY AND ASTROPHYSICS↗

Temperature and temporal dependence of the optical response for a radiochromic dosimeter

Purpose: Both temporal and thermal dependencies of the dose response have been observed in radiochromic dosimeters. As these dependencies may be influenced by the dose level, the present study investigates the temperature dependence during irradiation and the temporal change of the optical response following irradiation of radiochromic dosimeters at a range of doses. Methods: Cuvette samples of the PRESAGE Trade-Mark-Sign radiochromic dosimeter were irradiated within a dose range of 0-10 Gy at irradiation temperatures within 5-35 Degree-Sign C and postirradiation storage within 6-30 Degree-Sign C. The optical response due to irradiation was measured using a standard spectrophotometer and the data were analyzed in terms of thermal and temporal change. Results: The initial dose response was linear over the applied dose range independent of irradiation temperature. However, the optical response to a specific dose increased exponentially with irradiation temperature corresponding to an activation energy of 0.114 {+-} 0.007 eV. The temporal change in dose response after irradiation consisted of an offset, an auto-oxidation rate with activation energy 0.84 {+-} 0.03 eV, and an initial exponential increase in optical response (1.6 {+-} 0.2 eV) followed by an exponential decrease in optical response (0.98 {+-} 0.08 eV). These contributions depended on both storage temperature and the dose given, leading to a nonlinear dose response with time at low storage temperatures and a high auto-oxidation rate at high storage temperatures. Conclusions: Thermal equilibration is important to the radiochromic dosimeter investigated due to an exponential change in dose response with irradiation temperature and a considerable postirradiation temporal change in response. For the dosimeter version investigated in this study, a compromise in storage temperature has to be made between increasing the nonlinearity of the dose response with time and inducing a high auto-oxidation rate.

36 MATERIALS SCIENCE↗

SU-E-J-224: Multimodality Segmentation of Head and Neck Tumors

Purpose: Develop an algorithm that is able to automatically segment tumor volume in Head and Neck cancer by integrating information from CT, PET and MR imaging simultaneously. Methods: Twenty three patients that were recruited under an adaptive radiotherapy protocol had MR, CT and PET/CT scans within 2 months prior to start of radiotherapy. The patients had unresectable disease and were treated either with chemoradiotherapy or radiation therapy alone. Using the Velocity software, the PET/CT and MR (T1 weighted+contrast) scans were registered to the planning CT using deformable and rigid registration respectively. The PET and MR images were then resampled according to the registration to match the planning CT. The resampled images, together with the planning CT, were fed into a multi-channel segmentation algorithm, which is based on Gaussian mixture models and solved with the expectation-maximization algorithm and Markov random fields. A rectangular region of interest (ROI) was manually placed to identify the tumor area and facilitate the segmentation process. The auto-segmented tumor contours were compared with the gross tumor volume (GTV) manually defined by the physician. The volume difference and Dice similarity coefficient (DSC) between the manual and autosegmented GTV contours were calculated as the quantitative evaluation metrics. Results: The multimodality segmentation algorithm was applied to all 23 patients. The volumes of the auto-segmented GTV ranged from 18.4cc to 32.8cc. The average (range) volume difference between the manual and auto-segmented GTV was −42% (−32.8%–63.8%). The average DSC value was 0.62, ranging from 0.39 to 0.78. Conclusion: An algorithm for the automated definition of tumor volume using multiple imaging modalities simultaneously was successfully developed and implemented for Head and Neck cancer. This development along with more accurate registration algorithms can aid physicians in the efforts to interpret the multitude of imaging information available in radiotherapy today. This project was supported by a grant by Varian Medical Systems.

60 APPLIED LIFE SCIENCES↗

SU-E-J-131: Augmenting Atlas-Based Segmentation by Incorporating Image Features Proximal to the Atlas Contours

Purpose: For facilitating the current automatic segmentation, in this work we propose a narrow-shell strategy to enhance the information of each contour in the library and to improve the accuracy of the exiting atlas-based approach. Methods: In setting up an atlas-based library, we include not only the coordinates of contour points, but also the image features adjacent to the contour. 139 planning CT scans with normal appearing livers obtained during their radiotherapy treatment planning were used to construct the library. The CT images within the library were registered each other using affine registration. A nonlinear narrow shell with the regional thickness determined by the distance between two vertices alongside the contour. The narrow shell was automatically constructed both inside and outside of the liver contours. The common image features within narrow shell between a new case and a library case were first selected by a Speed-up Robust Features (SURF) strategy. A deformable registration was then performed using a thin plate splines (TPS) technique. The contour associated with the library case was propagated automatically onto the images of the new patient by exploiting the deformation field vectors. The liver contour was finally obtained by employing level set based energy function within the narrow shell. The performance of the proposed method was evaluated by comparing quantitatively the auto-segmentation results with that delineated by a physician. Results: Application of the technique to 30 liver cases suggested that the technique was capable of reliably segment organs such as the liver with little human intervention. Compared with the manual segmentation results by a physician, the average and discrepancies of the volumetric overlap percentage (VOP) was found to be 92.43%+2.14%. Conclusion: Incorporation of image features into the library contours improves the currently available atlas-based auto-contouring techniques and provides a clinically practical solution for auto-segmentation. This work is supported by NIH/NIBIB (1R01-EB016777), National Natural Science Foundation of China (No.61471226 and No.61201441), Research funding from Shandong Province (No.BS2012DX038 and No.J12LN23), and Research funding from Jinan City (No.201401221 and No.20120109)

60 APPLIED LIFE SCIENCES↗

WE-AB-BRA-09: Sensitivity of Plan Re-Optimization to Errors in Deformable Image Registration in Online Adaptive Image-Guided Radiation Therapy

Purpose: Online adaptive therapy (ART) relies on auto-contouring using deformable image registration (DIR). DIR’s inherent uncertainties require user intervention and manual edits while the patient is on the table. We investigated the dosimetric impact of DIR errors on the quality of re-optimized plans, and used the findings to establish regions for focusing manual edits to where DIR errors can Result in clinically relevant dose differences. Methods: Our clinical implementation of online adaptive MR-IGRT involves using DIR to transfer contours from CT to daily MR, followed by a physicians’ edits. The plan is then re-optimized to meet the organs at risk (OARs) constraints. Re-optimized abdomen and pelvis plans generated based on physician edited OARs were selected as the baseline for evaluation. Plans were then re-optimized on auto-deformed contours with manual edits limited to pre-defined uniform rings (0 to 5cm) around the PTV. A 0cm ring indicates that the auto-deformed OARs were used without editing. The magnitude of the variations caused by the non-deterministic optimizer was quantified by repeat re-optimizations on the same geometry to determine the mean and standard deviation (STD). For each re-optimized plan, various volumetric parameters for the PTV, the OARs were extracted along with DVH and isodose evaluation. A plan was deemed acceptable if the variation from the baseline plan was within one STD. Results: Initial results show that for abdomen and pancreas cases, a minimum of 5cm margin around the PTV is required for contour corrections, while for pelvic and liver cases a 2–3 cm margin is sufficient. Conclusion: Focusing manual contour edits to regions of dosimetric relevance can reduce contouring time in the online ART process while maintaining a clinically comparable plan. Future work will further refine the contouring region by evaluating the path along the beams, dose gradients near the target and OAR dose metrics.

60 APPLIED LIFE SCIENCES↗

SU-C-BRA-05: Delineating High-Dose Clinical Target Volumes for Head and Neck Tumors Using Machine Learning Algorithms

Purpose: To develop and test population-based machine learning algorithms for delineating high-dose clinical target volumes (CTVs) in H&N tumors. Automating and standardizing the contouring of CTVs can reduce both physician contouring time and inter-physician variability, which is one of the largest sources of uncertainty in H&N radiotherapy. Methods: Twenty-five node-negative patients treated with definitive radiotherapy were selected (6 right base of tongue, 11 left and 9 right tonsil). All patients had GTV and CTVs manually contoured by an experienced radiation oncologist prior to treatment. This contouring process, which is driven by anatomical, pathological, and patient specific information, typically results in non-uniform margin expansions about the GTV. Therefore, we tested two methods to delineate high-dose CTV given a manually-contoured GTV: (1) regression-support vector machines(SVM) and (2) classification-SVM. These models were trained and tested on each patient group using leave-one-out cross-validation. The volume difference(VD) and Dice similarity coefficient(DSC) between the manual and auto-contoured CTV were calculated to evaluate the results. Distances from GTV-to-CTV were computed about each patient’s GTV and these distances, in addition to distances from GTV to surrounding anatomy in the expansion direction, were utilized in the regression-SVM method. The classification-SVM method used categorical voxel-information (GTV, selected anatomical structures, else) from a 3×3×3cm3 ROI centered about the voxel to classify voxels as CTV. Results: Volumes for the auto-contoured CTVs ranged from 17.1 to 149.1cc and 17.4 to 151.9cc; the average(range) VD between manual and auto-contoured CTV were 0.93 (0.48–1.59) and 1.16(0.48–1.97); while average(range) DSC values were 0.75(0.59–0.88) and 0.74(0.59–0.81) for the regression-SVM and classification-SVM methods, respectively. Conclusion: We developed two novel machine learning methods to delineate high-dose CTV for H&N patients. Both methods showed promising results that hint to a solution to the standardization of the contouring process of clinical target volumes. Varian Medical Systems grant.

60 APPLIED LIFE SCIENCES↗

SU-C-BRA-06: Automatic Brain Tumor Segmentation for Stereotactic Radiosurgery Applications

Purpose: Stereotactic radiosurgery (SRS), which delivers a potent dose of highly conformal radiation to the target in a single fraction, requires accurate tumor delineation for treatment planning. We present an automatic segmentation strategy, that synergizes intensity histogram thresholding, super-voxel clustering, and level-set based contour evolving methods to efficiently and accurately delineate SRS brain tumors on contrast-enhance T1-weighted (T1c) Magnetic Resonance Images (MRI). Methods: The developed auto-segmentation strategy consists of three major steps. Firstly, tumor sites are localized through 2D slice intensity histogram scanning. Then, super voxels are obtained through clustering the corresponding voxels in 3D with reference to the similarity metrics composited from spatial distance and intensity difference. The combination of the above two could generate the initial contour surface. Finally, a localized region active contour model is utilized to evolve the surface to achieve the accurate delineation of the tumors. The developed method was evaluated on numerical phantom data, synthetic BRATS (Multimodal Brain Tumor Image Segmentation challenge) data, and clinical patients’ data. The auto-segmentation results were quantitatively evaluated by comparing to ground truths with both volume and surface similarity metrics. Results: DICE coefficient (DC) was performed as a quantitative metric to evaluate the auto-segmentation in the numerical phantom with 8 tumors. DCs are 0.999±0.001 without noise, 0.969±0.065 with Rician noise and 0.976±0.038 with Gaussian noise. DC, NMI (Normalized Mutual Information), SSIM (Structural Similarity) and Hausdorff distance (HD) were calculated as the metrics for the BRATS and patients’ data. Assessment of BRATS data across 25 tumor segmentation yield DC 0.886±0.078, NMI 0.817±0.108, SSIM 0.997±0.002, and HD 6.483±4.079mm. Evaluation on 8 patients with total 14 tumor sites yield DC 0.872±0.070, NMI 0.824±0.078, SSIM 0.999±0.001, and HD 5.926±6.141mm. Conclusion: The developed automatic segmentation strategy, which yields accurate brain tumor delineation in evaluation cases, is promising for its application in SRS treatment planning.

60 APPLIED LIFE SCIENCES↗

SU-F-T-94: Plan2pdf - a Software Tool for Automatic Plan Report for Philips Pinnacle TPS

Purpose: To implement an automatic electronic PDF plan reporting tool for Philips Pinnacle treatment planning system (TPS) Methods: An electronic treatment plan reporting software is developed by us to enable fully automatic PDF report from Pinnacle TPS to external EMR programs such as MOSAIQ. The tool is named “plan2pdf”. plan2pdf is implemented using Pinnacle scripts, Java and UNIX shell scripts, without any external program needed. plan2pdf supports full auto-mode and manual mode reporting. In full auto-mode, with a single mouse click, plan2pdf will generate a detailed Pinnacle plan report in PDF format, which includes customizable cover page, Pinnacle plan summary, orthogonal views through each plan POI and maximum dose point, DRR for each beam, serial transverse views captured throughout the dose grid at a user specified interval, DVH and scorecard windows. The final PDF report is also automatically bookmarked for each section above for convenient plan review. The final PDF report can either be saved on a user specified folder on Pinnacle, or it can be automatically exported to an EMR import folder via a user configured FTP service. In manual capture mode, plan2pdf allows users to capture any Pinnacle plan by full screen, individual window or rectangular ROI drawn on screen. Furthermore, to avoid possible patients’ plan mix-up during auto-mode reporting, a user conflict check feature is included in plan2pdf: it prompts user to wait if another patient is being exported by plan2pdf by another user. Results: plan2pdf is tested extensively and successfully at our institution consists of 5 centers, 15 dosimetrists and 10 physicists, running Pinnacle version 9.10 on Enterprise servers. Conclusion: plan2pdf provides a highly efficient, user friendly and clinical proven platform for all Philips Pinnacle users, to generate a detailed plan report in PDF format for external EMR systems.

60 APPLIED LIFE SCIENCES↗

TU-H-CAMPUS-JeP1-02: Fully Automatic Verification of Automatically Contoured Normal Tissues in the Head and Neck

Purpose: To investigate and validate the use of an independent deformable-based contouring algorithm for automatic verification of auto-contoured structures in the head and neck towards fully automated treatment planning. Methods: Two independent automatic contouring algorithms [(1) Eclipse’s Smart Segmentation followed by pixel-wise majority voting, (2) an in-house multi-atlas based method] were used to create contours of 6 normal structures of 10 head-and-neck patients. After rating by a radiation oncologist, the higher performing algorithm was selected as the primary contouring method, the other used for automatic verification of the primary. To determine the ability of the verification algorithm to detect incorrect contours, contours from the primary method were shifted from 0.5 to 2cm. Using a logit model the structure-specific minimum detectable shift was identified. The models were then applied to a set of twenty different patients and the sensitivity and specificity of the models verified. Results: Per physician rating, the multi-atlas method (4.8/5 point scale, with 3 rated as generally acceptable for planning purposes) was selected as primary and the Eclipse-based method (3.5/5) for verification. Mean distance to agreement and true positive rate were selected as covariates in an optimized logit model. These models, when applied to a group of twenty different patients, indicated that shifts could be detected at 0.5cm (brain), 0.75cm (mandible, cord), 1cm (brainstem, cochlea), or 1.25cm (parotid), with sensitivity and specificity greater than 0.95. If sensitivity and specificity constraints are reduced to 0.9, detectable shifts of mandible and brainstem were reduced by 0.25cm. These shifts represent additional safety margins which might be considered if auto-contours are used for automatic treatment planning without physician review. Conclusion: Automatically contoured structures can be automatically verified. This fully automated process could be used to flag auto-contours for special review or used with safety margins in a fully automatic treatment planning system.

60 APPLIED LIFE SCIENCES↗

TH-CD-206-05: Machine-Learning Based Segmentation of Organs at Risks for Head and Neck Radiotherapy Planning

Purpose: Accurate and efficient delineation of tumor target and organs-at-risks is essential for the success of radiotherapy. In reality, despite of decades of intense research efforts, auto-segmentation has not yet become clinical practice. In this study, we present, for the first time, a deep learning-based classification algorithm for autonomous segmentation in head and neck (HaN) treatment planning. Methods: Fifteen HN datasets of CT, MR and PET images with manual annotation of organs-at-risk (OARs) including spinal cord, brainstem, optic nerves, chiasm, eyes, mandible, tongue, parotid glands were collected and saved in a library of plans. We also have ten super-resolution MR images of the tongue area, where the genioglossus and inferior longitudinalis tongue muscles are defined as organs of interest. We applied the concepts of random forest- and deep learning-based object classification for automated image annotation with the aim of using machine learning to facilitate head and neck radiotherapy planning process. In this new paradigm of segmentation, random forests were used for landmark-assisted segmentation of super-resolution MR images. Alternatively to auto-segmentation with random forest-based landmark detection, deep convolutional neural networks were developed for voxel-wise segmentation of OARs in single and multi-modal images. The network consisted of three pairs of convolution and pooing layer, one RuLU layer and a softmax layer. Results: We present a comprehensive study on using machine learning concepts for auto-segmentation of OARs and tongue muscles for the HaN radiotherapy planning. An accuracy of 81.8% in terms of Dice coefficient was achieved for segmentation of genioglossus and inferior longitudinalis tongue muscles. Preliminary results of OARs regimentation also indicate that deep-learning afforded an unprecedented opportunities to improve the accuracy and robustness of radiotherapy planning. Conclusion: A novel machine learning framework has been developed for image annotation and structure segmentation. Our results indicate the great potential of deep learning in radiotherapy treatment planning.

60 APPLIED LIFE SCIENCES↗

TH-EF-207A-06: High-Resolution Optical-CT/ECT Imaging of Unstained Mice Femur, Brain, Spleen, and Tumor

Purpose: Optical transmission and emission computed tomography (optical-CT/ECT) provides high-resolution 3D attenuation and emission maps in unsectioned large (∼1cm{sup 3}) ex vivo tissue samples at a resolution of 12.9µm{sup 3} per voxel. Here we apply optical-CT/ECT to investigate high-resolution structure and auto-fluorescence in a range of optically cleared mice organs, including, for the first time, mouse bone (femur), opening the potential for study of bone metastasis and bone-mediated immune response. Methods: Three BALBc mice containing 4T1 flank tumors were sacrificed to obtain spleen, brain, tumor, and femur. Tissues were washed in 4% PFA, fixed in EtOH solution (for 5, 10, 10, and 2 days respectively), and then optically cleared for 3 days in BABBs. The femur was also placed in 0.25M aqueous EDTA for 15–30 days to remove calcium. Optical-CT/ECT attenuation and emission maps at 633nm (the latter using 530nm excitation light) were obtained for all samples. Bi-telecentric optical-CT was compared side-by-side with conventional optical projection tomography (OPT) imaging to evaluate imaging capability of these two rival techniques. Results: Auto-fluorescence mapping of femurs reveals vasculatures and fluorescence heterogeneity. High signals (A.U.=10) are reported in the medullary cavity but not in the cortical bone (A.U.=1). The brain strongly and uniform auto-fluoresces (A.U.=5). Thick, optically dense organs such as the spleen and the tumor (0.12, 0.46OD/mm) are reconstructed at depth without significant loss of resolution, which we attribute to the bi-telecentric optics of optical-CT. The attenuation map of tumor reveals vasculature, attenuation heterogeneity, and possibly necrotic tissue. Conclusion: We demonstrate the feasibility of optical-CT/ECT imaging of un-sectioned mice bones (femurs) and spleen with high resolution. This result, and the characterization of unstained organs, are important steps enabling future studies involving optical-CT/ECT applied to study metastasis and immunologic responses via fluorescence staining.

60 APPLIED LIFE SCIENCES↗

Dual-Zone Solar Control Indoor Shade: Demonstration at the Ronald V. Dellums Federal Building and U.S. Courthouse, 1301 Clay Street, Oakland, California and the Lawrence Berkeley National Laboratory Advanced Windows Testbed, Berkeley, California

Standard window shades are used to control sunlight, glare, window heat gains, and daylight. The challenge with conventional shades is that once the shades are lowered to reduce discomfort, daylight is reduced, which, in turn, increases the need for electric lighting. View is also obstructed, reducing the occupants’ connection to the outdoors. The strategy of subdividing the window wall and admitting daylight from above and controlling sunlight and view through the lower aperture can yield significant energy and peak demand savings over conventional shading systems, and improve the daylight quality of the indoor environment. This study evaluated a dual zone solar control (DZSC) shade with an upper, automated, horizontal, venetian blind with concave-up, matte white slats and a lower, manually operated, roller shade made of a transparent, reflective film. The automatic controls were designed to adjust the upper slat angle to maximize daylight to the room interior. The lower shade, when lowered, provided an unobstructed view to the outdoors. For the qualitative analysis, the DZSC shades were installed on windows in the southeast half of one floor in the Ronald V. Dellums Federal Building and U.S. Courthouse or hereafter referred to as the Oakland Federal Building (OFB) in Oakland, California. The position of the DZSC shades were observed and recorded periodically. Surveys were issued to occupants in the private offices and open plan workstations affected by the shades. For the quantitative analysis, the DZSC shades were installed in the south-facing windows of thermally isolated, fully instrumented, test chambers in the Advanced Windows Testbed at the Lawrence Berkeley National Laboratory (LBNL) in Berkeley, California. Measurements of window heat loads, daylight illuminance and environmental conditions related to visual and thermal comfort were collected over a six-month, solstice-to-solstice period. The performance of the DZSC shade was compared to fabric roller shades and venetian blinds since these are the two most common types of shades used in both commercial and Federal buildings. Several DZSC design permutations were evaluated: 1) manual or automated control of the upper louver system (“man” or “auto”) and 2) tinted or silvered reflective film for the lower shade (“Gray-Gray (GG)” or “Gray-Silver (GS)”1). The automated, reflective shade (auto-GS) reduced dimmable fluorescent lighting energy use by 51% and window cooling load by 13-15% compared to a conventional fabric roller shade with dual-pane, spectrally selective, low-e windows during the summer period (winter period was not measured). Qualitative analysis revealed that the auto-GS shade was preferred by 80% of the survey respondents compared to the existing vertical fabric blinds with single-pane, tinted windows in the occupied building. Survey results indicated that visual discomfort was reduced and thermal discomfort was also reduced during warm or hot weather. Occupants expressed greater satisfaction with the unobstructed view. Installation and commissioning of the DZSC shades was fairly straightforward with some minor problems due to inadequate installation instructions and improper wiring. A few occupants expressed dissatisfaction with the automatic controls for the upper louvers due to visual discomfort, erroneous control, and motor noise.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Terra-Populus v0.1: A Python Library for LandScan High-Definition Population Analysis and Modeling

The terra-populus library is designed for use by the LandScan HD technical team, offering a streamlined set of tools for generating and updating LandScan HD datasets from foundational building-level data, referred to as 'molecules,' provided by the building-level attribution team. This document serves as the primary technical documentation for terra-populus. Version 0.1 of the library includes the core modeling components necessary for LandScan HD production. It enables the generation of the LandScan HD Baseline dataset as well as corresponding confidence measures for the occupancy rates used. Parameters have been included for incorporating damaged building indicators and changes in population, to faciliate the creation of rapid updates for LandScan HD. Future iterations of terra-populus will introduce tools for creating a confidence index, and quantifying and propagating uncertainty, facilitating the creation of probabilistic LandScan HD outputs. This report provides an overview of the tools available in the library and the corresponding code implementations. One of the key advancements implemented in terra-populus is a redefinition of the atomic modeling unit for LandScan HD. Traditionally, the LandScan HD vector analytical framework has generated population estimates at the building sub-component (molecule) level. However, terra-populus adopts a building-level modeling approach. This shift is an operational decision aimed at aligning LandScan HD outputs with confidence measures, which are computed and validated at the building level (confidence measures are not included in this version of terra-populus, aside from those associated with the occupancy rates). Additional advancements to the LandScan HD modeling, as implemented by terra-populus, include a minimum population value parameter and an auto assignment of building floor counts. The population minimum value was implemented to prevent buildings and subsequent LandScan HD pixels that contained small values that may not rasterize in production. An 'auto' value has been included as a method for dealing with buildings lacking floor count information, where it is the average floor count of all other buildings with a residential building use type tag. The logic behind this is to remain consistent with the current logic employed for dealing with building use type null instances, where a null use type is defaulted to residential since it is the most common building type. The auto logic is intended to apply the most common building floor count of the most common type of buildings. The tools provided in terra-populus represent a significant step forward in improving the efficiency, reproducibility, and transparency of the LandScan HD modeling process. As the library evolves, it will continue to serve as a foundational resource for high-resolution population modeling.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗