The proceedings contain 39 papers. The topics discussed include: single view homography estimation for an inclined textured planar surface: overcoming the inverse and ill-posed challenge!;MMAG: mutually motivated atte...
ISBN:
(纸本)9798400716256
The proceedings contain 39 papers. The topics discussed include: single view homography estimation for an inclined textured planar surface: overcoming the inverse and ill-posed challenge!;MMAG: mutually motivated attention gates for simultaneous extraction of contextual and spatial information from a monocular image;automatic assessment of communication skill in real-world job interviews: a comparative study using deep learning and domain adaptation;an efficient motor imagery classification framework using sparse brain connectivity and class-consistent dictionary learning from electroencephalogram signals;mandala as computational art: vectorization and beyond;degradation aware multi-scale approach to no reference image quality assessment;dual stage semantic information based generative adversarial network for image super-resolution;knowledge distillation with ensemble calibration;and a novel framework for robust fingerprint representations using deep convolution network with attention mechanism.
the proceedings contain 59 papers. the topics discussed include: interpreting intrinsic image decomposition using concept activations;quaternion factorized simulated exposure fusion;learning from multiple datasets for...
ISBN:
(纸本)9781450398237
the proceedings contain 59 papers. the topics discussed include: interpreting intrinsic image decomposition using concept activations;quaternion factorized simulated exposure fusion;learning from multiple datasets for recognizing human actions;topological shape matching using multi-dimensional Reeb graphs;convolutional ensembling based few-shot defect detection technique;performance, trust, or both? COVID-19 diagnosis and prognosis using deep ensemble transfer learning on x-ray images;Alzheimer’s severity classification using transfer learning and residual separable convolution network;detecting coronavirus (COVID–19) disease cues from chest radiography images;posture guided human action recognition for fitness applications;towards robust handwritten text recognition with on-the-fly user participation;low resource degraded quality document image binarization – domain adaptation is the way;a globally-connected and trainable hierarchical fine-attention generative adversarial network based adversarial defense;and overcoming label noise for source-free unsupervised video domain adaptation.
The complete perception and segmentation of images has become a research hotspot in the field of computervision with the continuous development of the computervision industry. This image segmentation technique needs...
详细信息
ISBN:
(纸本)9798400713880
The complete perception and segmentation of images has become a research hotspot in the field of computervision with the continuous development of the computervision industry. This image segmentation technique needs the understanding of a scene which can neither figure out individual instances or different actors in the environment as well as the general background semantics. This newly emerging panopic segmentation problem in recent years is solved in this paper with a lightweight and efficient panoptic segmentation network, LEPSNet, which is based on Efficientnet. We designed this architecture based on a Two way PAN, with the ability of efficiently encoding and fusing semantically rich multi-scale features as a backbone network. A new semantic segmentation head is designed to capture fine context information in the background. Proposed is an improved separable convolutional module, and applied to the whole network. This module also results in adding it to Mask R-CNN, making the instance segmentation head. Experiments have demonstrated that the architecture can achieve a level of performance leadership while keeping the model lightweight. In particular, we achieved 65.0 PQ on Cityscapes and 50.5 PQ on indian Driving Dataset (IDD).
One of the foremost requisite for human perception and computervision task is to get an image with all objects in focus. The image fusion process, as one of the solutions, allows getting a clear fused image from seve...
详细信息
ISBN:
(纸本)9781479915880
One of the foremost requisite for human perception and computervision task is to get an image with all objects in focus. The image fusion process, as one of the solutions, allows getting a clear fused image from several images acquired with different focus levels of a scene. In this paper, a novel framework for multi-focus image fusion is proposed, which is computationally simple since it realizes only in the spatial domain. The proposed framework is based on the fractal dimensions of the images into the fusion process. The extensive experiments on different multifocus image sets demonstrate that it is consistently superior to the conventional image fusion methods in terms of visual and quantitative evaluations.
In all imageprocessing applications, it is important to extract the appropriate information from an image. But often the captured image is not clear enough to give the required information due to the imaging environm...
详细信息
ISBN:
(纸本)9781479915880
In all imageprocessing applications, it is important to extract the appropriate information from an image. But often the captured image is not clear enough to give the required information due to the imaging environment. Thus, it is essential to enhance the clarity of the image by some post-processing techniques. image deblurring is one of such techniques to remove the blurry effect of the captured image. This paper looks into this problem in a different way, where the deblurring of an image is addressed by solving image super-resolution problem. The blurred image is first down-sampled and then it is fed to the super-resolution framework to produce the deblurred high resolution image. In addition, the proposed approach states the requirement of edge preservation in the problem. The experimental results are comparable with the existing image deblurring algorithms.
A perceptual video hashing function maps the perceptual content of a video into a fixed-length binary string called the perceptual hash. Perceptual hashing is a promising solution to the content-identification and the...
详细信息
ISBN:
(纸本)9781479915880
A perceptual video hashing function maps the perceptual content of a video into a fixed-length binary string called the perceptual hash. Perceptual hashing is a promising solution to the content-identification and the content-authentication problems. The projections of image and video data onto a subspace have been exploited in the literature to get a compact hash function. We propose a new perceptual video hashing algorithm based on the Achlioptas's random projections. Simulation results show that the proposed perceptual hash function is robust to common signal and imageprocessing attacks.
In this paper, we present a fast and efficient algorithm for regularization and resampling of triangular meshes generated by 3D reconstruction methods such as stereoscopy, laser scanning etc. We also present a scheme ...
详细信息
ISBN:
(纸本)9781479915880
In this paper, we present a fast and efficient algorithm for regularization and resampling of triangular meshes generated by 3D reconstruction methods such as stereoscopy, laser scanning etc. We also present a scheme for efficient parallel implementation of the proposed algorithm and the time gain with increasing number of processor cores.
Approximate Nearest-Neighbour Field has been an area of interest in recent research for a wide variety of topics in graphics and multimedia community. Medical imageprocessing is a relatively unaffected field by these...
详细信息
ISBN:
(纸本)9781479915880
Approximate Nearest-Neighbour Field has been an area of interest in recent research for a wide variety of topics in graphics and multimedia community. Medical imageprocessing is a relatively unaffected field by these developments in ANNF computations, brought about by various extremely efficient algorithms like PatchMatch. In this paper, we use Generalized PatchMatch for Optic Disk detection, in retinal images, and show that by making use of efficient ANNF computations we are able to generate results with 98% accuracy with an average time of 0.5 sec. This is significantly faster than conventional Optic Disk detection methods, which average at 95-97% accuracy with 3-5 sec average computation time.
暂无评论