Hardware architecture and optimization of sliding window based pedestrian detection on FPGA for high resolution images by varying local features

Author(s):  
Asim Khan ◽  
Muhammad Umar Karim Khan ◽  
Muhammad Bilal ◽  
Chong-Min Kyung
Electronics ◽  
2019 ◽  
Vol 8 (11) ◽  
pp. 1370 ◽  
Author(s):  
Tingzhu Sun ◽  
Weidong Fang ◽  
Wei Chen ◽  
Yanxin Yao ◽  
Fangming Bi ◽  
...  

Although image inpainting based on the generated adversarial network (GAN) has made great breakthroughs in accuracy and speed in recent years, they can only process low-resolution images because of memory limitations and difficulty in training. For high-resolution images, the inpainted regions become blurred and the unpleasant boundaries become visible. Based on the current advanced image generation network, we proposed a novel high-resolution image inpainting method based on multi-scale neural network. This method is a two-stage network including content reconstruction and texture detail restoration. After holding the visually believable fuzzy texture, we further restore the finer details to produce a smoother, clearer, and more coherent inpainting result. Then we propose a special application scene of image inpainting, that is, to delete the redundant pedestrians in the image and ensure the reality of background restoration. It involves pedestrian detection, identifying redundant pedestrians and filling in them with the seemingly correct content. To improve the accuracy of image inpainting in the application scene, we proposed a new mask dataset, which collected the characters in COCO dataset as a mask. Finally, we evaluated our method on COCO and VOC dataset. the experimental results show that our method can produce clearer and more coherent inpainting results, especially for high-resolution images, and the proposed mask dataset can produce better inpainting results in the special application scene.


2021 ◽  
Vol 10 (12) ◽  
pp. 813
Author(s):  
Osmar Luiz Ferreira de Carvalho ◽  
Rebeca dos Santos de Moura ◽  
Anesmar Olino de Albuquerque ◽  
Pablo Pozzobon de Bem ◽  
Rubens de Castro Pereira ◽  
...  

Misappropriation of public lands is an ongoing government concern. In Brazil, the beach zone is public property, but many private establishments use it for economic purposes, requiring constant inspection. Among the undue targets, the individual mapping of straw beach umbrellas (SBUs) attached to the sand is a great challenge due to their small size, high presence, and agglutinated appearance. This study aims to automatically detect and count SBUs on public beaches using high-resolution images and instance segmentation, obtaining pixel-wise semantic information and individual object detection. This study is the first instance segmentation application on coastal areas and the first using WorldView-3 (WV-3) images. We used the Mask-RCNN with some modifications: (a) multispectral input for the WorldView3 imagery (eight channels), (b) improved the sliding window algorithm for large image classification, and (c) comparison of different image resizing ratios to improve small object detection since the SBUs are small objects (<322 pixels) even using high-resolution images (31 cm). The accuracy analysis used standard COCO metrics considering the original image and three scale ratios (2×, 4×, and 8× resolution increase). The average precision (AP) results increased proportionally to the image resolution: 30.49% (original image), 48.24% (2×), 53.45% (4×), and 58.11% (8×). The 8× model presented 94% AP50, classifying nearly all SBUs correctly. Moreover, the improved sliding window approach enables the classification of large areas providing automatic counting and estimating the size of the objects, proving to be effective for inspecting large coastal areas and providing insightful information for public managers. This remote sensing application impacts the inspection cost, tribute, and environmental conditions.


1994 ◽  
Vol 144 ◽  
pp. 541-547
Author(s):  
J. Sýkora ◽  
J. Rybák ◽  
P. Ambrož

AbstractHigh resolution images, obtained during July 11, 1991 total solar eclipse, allowed us to estimate the degree of solar corona polarization in the light of FeXIV 530.3 nm emission line and in the white light, as well. Very preliminary analysis reveals remarkable differences in the degree of polarization for both sets of data, particularly as for level of polarization and its distribution around the Sun’s limb.


Author(s):  
Etienne de Harven

Biological ultrastructures have been extensively studied with the scanning electron microscope (SEM) for the past 12 years mainly because this instrument offers accurate and reproducible high resolution images of cell shapes, provided the cells are dried in ways which will spare them the damage which would be caused by air drying. This can be achieved by several techniques among which the critical point drying technique of T. Anderson has been, by far, the most reproducibly successful. Many biologists, however, have been interpreting SEM micrographs in terms of an exclusive secondary electron imaging (SEI) process in which the resolution is primarily limited by the spot size of the primary incident beam. in fact, this is not the case since it appears that high resolution, even on uncoated samples, is probably compromised by the emission of secondary electrons of much more complex origin.When an incident primary electron beam interacts with the surface of most biological samples, a large percentage of the electrons penetrate below the surface of the exposed cells.


Author(s):  
S. Saito ◽  
H. Todokoro ◽  
S. Nomura ◽  
T. Komoda

Field emission scanning electron microscope (FESEM) features extremely high resolution images, and offers many valuable information. But, for a specimen which gives low contrast images, lateral stripes appear in images. These stripes are resulted from signal fluctuations caused by probe current noises. In order to obtain good images without stripes, the fluctuations should be less than 1%, especially for low contrast images. For this purpose, the authors realized a noise compensator, and applied this to the FESEM.Fig. 1 shows an outline of FESEM equipped with a noise compensator. Two apertures are provided gust under the field emission gun.


Author(s):  
David C. Joy ◽  
Dennis M. Maher

High-resolution images of the surface topography of solid specimens can be obtained using the low-loss technique of Wells. If the specimen is placed inside a lens of the condenser/objective type, then it has been shown that the lens itself can be used to collect and filter the low-loss electrons. Since the probeforming lenses in TEM instruments fitted with scanning attachments are of this type, low-loss imaging should be possible.High-resolution, low-loss images have been obtained in a JEOL JEM 100B fitted with a scanning attachment and a thermal, fieldemission gun. No modifications were made to the instrument, but a wedge-shaped, specimen holder was made to fit the side-entry, goniometer stage. Thus the specimen is oriented initially at a glancing angle of about 30° to the beam direction. The instrument is set up in the conventional manner for STEM operation with all the lenses, including the projector, excited.


Author(s):  
M. Kelly ◽  
D.M. Bird

It is well known that strain fields can have a strong influence on the details of HREM images. This, for example, can cause problems in the analysis of edge-on interfaces between lattice mismatched materials. An interesting alternative to conventional HREM imaging has recently been advanced by Pennycook and co-workers where the intensity variation in the annular dark field (ADF) detector is monitored as a STEM probe is scanned across the specimen. It is believed that the observed atomic-resolution contrast is correlated with the intensity of the STEM probe at the atomic sites and the way in which this varies as the probe moves from cell to cell. As well as providing a directly interpretable high-resolution image, there are reasons for believing that ADF-STEM images may be less suseptible to strain than conventional HREM. This is because HREM images arise from the interference of several diffracted beams, each of which is governed by all the excited Bloch waves in the crystal.


Author(s):  
Kenneth H. Downing ◽  
Hu Meisheng ◽  
Hans-Rudolf Went ◽  
Michael A. O'Keefe

With current advances in electron microscope design, high resolution electron microscopy has become routine, and point resolutions of better than 2Å have been obtained in images of many inorganic crystals. Although this resolution is sufficient to resolve interatomic spacings, interpretation generally requires comparison of experimental images with calculations. Since the images are two-dimensional representations of projections of the full three-dimensional structure, information is invariably lost in the overlapping images of atoms at various heights. The technique of electron crystallography, in which information from several views of a crystal is combined, has been developed to obtain three-dimensional information on proteins. The resolution in images of proteins is severely limited by effects of radiation damage. In principle, atomic-resolution, 3D reconstructions should be obtainable from specimens that are resistant to damage. The most serious problem would appear to be in obtaining high-resolution images from areas that are thin enough that dynamical scattering effects can be ignored.


Sign in / Sign up

Export Citation Format

Share Document