Search NASA⌕ Search

SEARCH · Search NASA

Results for “Image Transforms”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS1982

Integrated optics in an electrically scanned imaging Fourier transform spectrometer

An efficient, lightweight and stable, Fourier transform spectrometer was developed. The mechanical slide mechanism needed to create a path difference was eliminated by the use of retro-reflecting mirrors in a monolithic interferometer assembly in which the mirrors are not at 90 degrees to the propagation vector of the radiation, but rather at a small angle. The resulting plane wave fronts create a double-sided inteferogram of the source irradiance distribution which is detected by a charge-coupled device image sensor array. The position of each CCD pixel in the array is an indication of the path difference between the two retro-reflecting mirrors in the monolithic optical structure. The Fourier transform of the signals generated by the image sensor provide the spectral irradiance distribution of the source. For imaging, the interferometer assembly scans the source of irradiation by moving the entire instrument, such as would occur if it was fixedly mounted to a moving platform, i.e., a spacecraft. During scanning, the entrace slot to the monolithic optical structure sends different pixels to corresponding interferograms detected by adjacent columns of pixels of the image sensor.

Breckinridge, James B.↗
NASA NTRS2009

Sequential Principal Component Analysis -An Optimal and Hardware-Implementable Transform for Image Compression

This paper presents the JPL-developed Sequential Principal Component Analysis (SPCA) algorithm for feature extraction / image compression, based on "dominant-term selection" unsupervised learning technique that requires an order-of-magnitude lesser computation and has simpler architecture compared to the state of the art gradient-descent techniques. This algorithm is inherently amenable to a compact, low power and high speed VLSI hardware embodiment. The paper compares the lossless image compression performance of the JPL's SPCA algorithm with the state of the art JPEG2000, widely used due to its simplified hardware implementability. JPEG2000 is not an optimal data compression technique because of its fixed transform characteristics, regardless of its data structure. On the other hand, conventional Principal Component Analysis based transform (PCA-transform) is a data-dependent-structure transform. However, it is not easy to implement the PCA in compact VLSI hardware, due to its highly computational and architectural complexity. In contrast, the JPL's "dominant-term selection" SPCA algorithm allows, for the first time, a compact, low-power hardware implementation of the powerful PCA algorithm. This paper presents a direct comparison of the JPL's SPCA versus JPEG2000, incorporating the Huffman and arithmetic coding for completeness of the data compression operation. The simulation results show that JPL's SPCA algorithm is superior as an optimal data-dependent-transform over the state of the art JPEG2000. When implemented in hardware, this technique is projected to be ideally suited to future NASA missions for autonomous on-board image data processing to improve the bandwidth of communication.

Duong, Tuan A.↗
NASA NTRS1993

High-speed joint-transform optical image correlator using GaAs/AlGaAs semi-insulating multiple quantum wells and diode lasers

We demonstrate a compact real-time optical image correlator using diode lasers and a semi-insulating GaAs/AlGaAs multiple-quantum-well (SI-MQW) device as the holographic element. With only 3 mW of power incident upon the SI-MQW device, the correlation is obtained in 1 microsec and can be erased in as fast as 2 microsec, which presents the possibility of a system capable of 3 x 10 exp 5 correlations/s. We also show that images can be stored for a controllable length of time (2-25 microsec), which presents possibilities for the use of the SI-MQW device as a data or image buffer memory.

Partovi, A.↗
NASA NTRS1995

Optical Logarithmic Transformation of Speckle Images with Bacteriorhodopsin Films

The application of logarithmic transformations to speckle images is sometimes desirable in converting the speckle noise distribution into an additive, constant-variance noise distribution. The optical transmission properties of some bacteriorhodopsin films are well suited to implement such a transformation optically in a parallel fashion. I present experimental results of the optical conversion of a speckle image into a transformed image with signal-independent noise statistics, using the real-time photochromic properties of bacteriorhodopsin. The original and transformed noise statistics are confirmed by histogram analysis.

Downie, John D.↗
DOE OSTI2025

Sequence length scaling in vision transformers for scientific images on frontier

Vision Transformers (ViTs) are pivotal for foundational models in scientific imagery, including Earth science applications, due to their capability to process large sequence lengths. While transformers for text have inspired scaling sequence lengths in ViTs, adapting these for ViTs introduces unique challenges. We develop distributed sequence parallelism for ViTs, enabling them to handle up to 1M tokens. Our approach, leveraging DeepSpeed-Ulysses and Long-Sequence-Segmentation with model sharding, is the first to apply sequence parallelism in ViT training, achieving a 94% batch scaling efficiency on 2,048 AMD-MI250X GPUs. Evaluating sequence parallelism in ViTs, particularly in models up to 10B parameters, highlighted substantial bottlenecks. We countered these with hybrid sequence, pipeline, and flash attention strategies, to scale beyond single GPU memory limits. Our method significantly enhances climate modeling accuracy by 20% in temperature predictions, marking the first training of a vision transformer model to convergence with a sequence length of 188K tokens, using full self-attention.

Tsaris, Aristeidis (aris) [ORNL] (ORCID:0000000277↗
NASA NTRS2011

Geometric registration of images by similarity transformation using two reference points

A method for registering a first image to a second image using a similarity transformation. The each image includes a plurality of pixels. The first image pixels are mapped to a set of first image coordinates and the second image pixels are mapped to a set of second image coordinates. The first image coordinates of two reference points in the first image are determined. The second image coordinates of these reference points in the second image are determined. A Cartesian translation of the set of second image coordinates is performed such that the second image coordinates of the first reference point match its first image coordinates. A similarity transformation of the translated set of second image coordinates is performed. This transformation scales and rotates the second image coordinates about the first reference point such that the second image coordinates of the second reference point match its first image coordinates.

Kang, Yong Q.↗
NASA NTRS1995

Subband/Transform MATLAB Functions For Processing Images

SUBTRANS software is package of routines implementing image-data-processing functions for use with MATLAB*(TM) software. Provides capability to transform image data with block transforms and to produce spatial-frequency subbands of transformed data. Functions cascaded to provide further decomposition into more subbands. Also used in image-data-compression systems. For example, transforms used to prepare data for lossy compression. Written for use in MATLAB mathematical-analysis environment.

Glover, D.↗
NASA NTRS1993

DCT quantization matrices visually optimized for individual images

This presentation describes how a vision model incorporating contrast sensitivity, contrast masking, and light adaptation is used to design visually optimal quantization matrices for Discrete Cosine Transform image compression. The Discrete Cosine Transform (DCT) underlies several image compression standards (JPEG, MPEG, H.261). The DCT is applied to 8x8 pixel blocks, and the resulting coefficients are quantized by division and rounding. The 8x8 'quantization matrix' of divisors determines the visual quality of the reconstructed image; the design of this matrix is left to the user. Since each DCT coefficient corresponds to a particular spatial frequency in a particular image region, each quantization error consists of a local increment or decrement in a particular frequency. After adjustments for contrast sensitivity, local light adaptation, and local contrast masking, this coefficient error can be converted to a just-noticeable-difference (jnd). The jnd's for different frequencies and image blocks can be pooled to yield a global perceptual error metric. With this metric, we can compute for each image the quantization matrix that minimizes the bit-rate for a given perceptual error, or perceptual error for a given bit-rate. Implementation of this system demonstrates its advantages over existing techniques. A unique feature of this scheme is that the quantization matrix is optimized for each individual image. This is compatible with the JPEG standard, which requires transmission of the quantization matrix.

Watson, Andrew B.↗
NASA NTRS1993

Subband/transform functions for image processing

Functions for image data processing written for use with the MATLAB(TM) software package are presented. These functions provide the capability to transform image data with block transformations (such as the Walsh Hadamard) and to produce spatial frequency subbands of the transformed data. Block transforms are equivalent to simple subband systems. The transform coefficients are reordered using a simple permutation to give subbands. The low frequency subband is a low resolution version of the original image, while the higher frequency subbands contain edge information. The transform functions can be cascaded to provide further decomposition into more subbands. If the cascade is applied to all four of the first stage subbands (in the case of a four band decomposition), then a uniform structure of sixteen bands is obtained. If the cascade is applied only to the low frequency subband, an octave structure of seven bands results. Functions for the inverse transforms are also given. These functions can be used for image data compression systems. The transforms do not in themselves produce data compression, but prepare the data for quantization and compression. Sample quantization functions for subbands are also given. A typical compression approach is to subband the image data, quantize it, then use statistical coding (e.g., run-length coding followed by Huffman coding) for compression. Contour plots of image data and subbanded data are shown.

Glover, Daniel↗
NASA NTRS1994

A Motionless Camera

Omniview, a motionless, noiseless, exceptionally versatile camera was developed for NASA as a receiving device for guiding space robots. The system can see in one direction and provide as many as four views simultaneously. Developed by Omniview, Inc. (formerly TRI) under a NASA Small Business Innovation Research (SBIR) grant, the system's image transformation electronics produce a real-time image from anywhere within a hemispherical field. Lens distortion is removed, and a corrected "flat" view appears on a monitor. Key elements are a high resolution charge coupled device (CCD), image correction circuitry and a microcomputer for image processing. The system can be adapted to existing installations. Applications include security and surveillance, teleconferencing, imaging, virtual reality, broadcast video and military operations. Omniview technology is now called IPIX. The company was founded in 1986 as TeleRobotics International, became Omniview in 1995, and changed its name to Interactive Pictures Corporation in 1997.

Source record↗
NASA NTRS1975

Gray-level transformations for interactive image enhancement

A gray-level transformation method suitable for interactive image enhancement was presented. It is shown that the well-known histogram equalization approach is a special case of this method. A technique for improving the uniformity of a histogram is also developed. Experimental results which illustrate the capabilities of both algorithms are described. Two proposals for implementing gray-level transformations in a real-time interactive image enhancement system are also presented.

Fittes, B. A.↗
NASA NTRS1987

A laboratory demonstration of high-resolution hard X-ray and gamma-ray imaging using Fourier-transform techniques

A laboratory imaging system has been developed to study the use of Fourier-transform techniques in high-resolution hard X-ray and gamma-ray imaging, with particular emphasis on possible applications to high-energy astronomy. Considerations for the design of a Fourier-transform imager and the instrumentation used in the laboratory studies is described. Several analysis methods for image reconstruction are discussed including the CLEAN algorithm and maximum entropy methods. Images obtained using these methods are presented.

Palmer, David↗
DOE OSTI2024

Adaptive Patching for High-resolution Image Segmentation with Transformers

Attention-based models are proliferating in the space of image analytics, including segmentation. The standard method of feeding images to transformer encoders is to divide the images into patches and then feed the patches to the model as a linear sequence of tokens. For high-resolution images, e.g. microscopic pathology images, the quadratic compute and memory cost prohibits the use of an attention-based model, if we are to use smaller patch sizes that are favorable in segmentation. The solution is to either use custom complex multi-resolution models or approximate attention schemes. We take inspiration from Adapative Mesh Refinement (AMR) methods in HPC by adaptively patching the images, as a pre-processing step, based on the image details to reduce the number of patches being fed to the model, by orders of magnitude. This method has a negligible overhead, and works seamlessly with any attention-based model, i.e. it is a pre-processing step that can be adopted by any attention-based model without friction. We demonstrate superior segmentation quality over SoTA segmentation models for realworld pathology datasets while gaining a geomean speedup of 6.9× for resolutions up to 64K2, on up to 2, 048 GPUs.

Zhang, Enzhi↗
NASA NTRS2006

Electro-Optical Imaging Fourier-Transform Spectrometer

An electro-optical (E-O) imaging Fourier-transform spectrometer (IFTS), now under development, is a prototype of improved imaging spectrometers to be used for hyperspectral imaging, especially in the infrared spectral region. Unlike both imaging and non-imaging traditional Fourier-transform spectrometers, the E-O IFTS does not contain any moving parts. Elimination of the moving parts and the associated actuator mechanisms and supporting structures would increase reliability while enabling reductions in size and mass, relative to traditional Fourier-transform spectrometers that offer equivalent capabilities. Elimination of moving parts would also eliminate the vibrations caused by the motions of those parts. Figure 1 schematically depicts a traditional Fourier-transform spectrometer, wherein a critical time delay is varied by translating one the mirrors of a Michelson interferometer. The time-dependent optical output is a periodic representation of the input spectrum. Data characterizing the input spectrum are generated through fast-Fourier-transform (FFT) post-processing of the output in conjunction with the varying time delay.

Chao, Tien-Hsin↗
NASA NTRS1975

Gray-level transformations for interactive image enhancement.

A gray-level transformation method suitable for interactive image enhancement is presented. It is shown that the well-known histogram equalization technique is a special case of this method. Experimental results which establish the superiority of the procedure over histogram equalization are also described.

Gonzalez, R. C.↗
NASA NTRS1977

Gray-level transformations for interactive image enhancement

A gray-level transformation method suitable for interactive image enhancement is presented. It is shown that the well-known histogram equalization technique is a special case of this method. Experimental results which illustrate the enhancement capabilities of the procedure are also described.

Gonzalez, R. C.↗