arXiv subjects

Shahram Shirani

Publications and source records attributed to Shahram Shirani.

At least 37 records · Page 2Linked to original sources

arXiv2024

Robust Multi-Modal Speech In-Painting: A Sequence-to-Sequence Approach

The process of reconstructing missing parts of speech audio from context is called speech in-painting. Human perception of speech is inherently multi-modal, involving both audio and visual (AV) cues. In this paper, we introduce and study a sequence-to-sequence (seq2seq) speech in-painting model that incorporates AV features. Our approach extends AV speech in-painting techniques to scenarios where both audio and visual data may be jointly corrupted. To achieve this, we employ a multi-modal training paradigm that boosts the robustness of our model across various conditions involving acoustic and visual distortions. This makes our distortion-aware model a plausible solution for real-world challenging environments. We compare our method with existing transformer-based and recurrent neural network-based models, which attempt to reconstruct missing speech gaps ranging from a few milliseconds to over a second. Our experimental results demonstrate that our novel seq2seq architecture outperforms the state-of-the-art transformer solution by 38.8% in terms of enhancing speech quality and 7.14% in terms of improving speech intelligibility. We exploit a multi-task learning framework that simultaneously performs lip-reading (transcribing video components to text) while reconstructing missing parts of the associated speech.

cs.MM↗

arXiv2024

A Parametric Rate-Distortion Model for Video Transcoding

Over the past two decades, the surge in video streaming applications has been fueled by the increasing accessibility of the internet and the growing demand for network video. As users with varying internet speeds and devices seek high-quality video, transcoding becomes essential for service providers. In this paper, we introduce a parametric rate-distortion (R-D) transcoding model. Our model excels at predicting transcoding distortion at various rates without the need for encoding the video. This model serves as a versatile tool that can be used to achieve visual quality improvement (in terms of PSNR) via trans-sizing. Moreover, we use our model to identify visually lossless and near-zero-slope bitrate ranges for an ingest video. Having this information allows us to adjust the transcoding target bitrate while introducing visually negligible quality degradations. By utilizing our model in this manner, quality improvements up to 2 dB and bitrate savings of up to 46% of the original target bitrate are possible. Experimental results demonstrate the efficacy of our model in video transcoding rate distortion prediction.

cs.MM↗

arXiv2023

Supervised Deep Learning for Content-Aware Image Retargeting with Fourier Convolutions

Image retargeting aims to alter the size of the image with attention to the contents. One of the main obstacles to training deep learning models for image retargeting is the need for a vast labeled dataset. Labeled datasets are unavailable for training deep learning models in the image retargeting tasks. As a result, we present a new supervised approach for training deep learning models. We use the original images as ground truth and create inputs for the model by resizing and cropping the original images. A second challenge is generating different image sizes in inference time. However, regular convolutional neural networks cannot generate images of different sizes than the input image. To address this issue, we introduced a new method for supervised learning. In our approach, a mask is generated to show the desired size and location of the object. Then the mask and the input image are fed to the network. Comparing image retargeting methods and our proposed method demonstrates the model's ability to produce high-quality retargeted images. Afterward, we compute the image quality assessment score for each output image based on different techniques and illustrate the effectiveness of our approach.

cs.CV↗

arXiv2023

Endoscopy Classification Model Using Swin Transformer and Saliency Map

Endoscopy is a valuable tool for the early diagnosis of colon cancer. However, it requires the expertise of endoscopists and is a time-consuming process. In this work, we propose a new multi-label classification method, which considers two aspects of learning approaches (local and global views) for endoscopic image classification. The model consists of a Swin transformer branch and a modified VGG16 model as a CNN branch. To help the learning process of the CNN branch, the model employs saliency maps and endoscopy images and concatenates them. The results demonstrate that this method performed well for endoscopic medical images by utilizing local and global features of the images. Furthermore, quantitative evaluations prove the proposed method's superiority over state-of-the-art works.

eess.IV↗

arXiv2023

SFI-Swin: Symmetric Face Inpainting with Swin Transformer by Distinctly Learning Face Components Distributions

Image inpainting consists of filling holes or missing parts of an image. Inpainting face images with symmetric characteristics is more challenging than inpainting a natural scene. None of the powerful existing models can fill out the missing parts of an image while considering the symmetry and homogeneity of the picture. Moreover, the metrics that assess a repaired face image quality cannot measure the preservation of symmetry between the rebuilt and existing parts of a face. In this paper, we intend to solve the symmetry problem in the face inpainting task by using multiple discriminators that check each face organ's reality separately and a transformer-based network. We also propose "symmetry concentration score" as a new metric for measuring the symmetry of a repaired face image. The quantitative and qualitative results show the superiority of our proposed method compared to some of the recently proposed algorithms in terms of the reality, symmetry, and homogeneity of the inpainted parts.

cs.CV↗

arXiv2022

Progressive Hologram Generation Based on Object Saliency

Computer-generated hologram (CGH) is promised to realize the next generation of 3D visual media with life-changing applications. However, one of the essential obstacles to this technology is the time-consuming hologram computation. Thus, facilitating the computation of the generated hologram is of significant importance in this area. We propose a progressive hologram generation based on object saliency using discrete wavelet transform. In our method, the object is decomposed into 3 levels of resolution using wavelet transform. Then, based on the saliency of the object, a progressive resolution hologram is generated. Our model generates a low resolution hologram for non-salient areas of an object and a high-resolution hologram for salient areas, thus reducing CGH generation time. We applied our method to a number of objects and show that salient parts are reconstructed with high quality while it helps the process to speed up. Finally, we compare the SSIM of the reconstructed objects.

cs.MM↗

arXiv2022

Dynamic-Pix2Pix: Noise Injected cGAN for Modeling Input and Target Domain Joint Distributions with Limited Training Data

Learning to translate images from a source to a target domain with applications such as converting simple line drawing to oil painting has attracted significant attention. The quality of translated images is directly related to two crucial issues. First, the consistency of the output distribution with that of the target is essential. Second, the generated output should have a high correlation with the input. Conditional Generative Adversarial Networks, cGANs, are the most common models for translating images. The performance of a cGAN drops when we use a limited training dataset. In this work, we increase the Pix2Pix (a form of cGAN) target distribution modeling ability with the help of dynamic neural network theory. Our model has two learning cycles. The model learns the correlation between input and ground truth in the first cycle. Then, the model's architecture is refined in the second cycle to learn the target distribution from noise input. These processes are executed in each iteration of the training procedure. Helping the cGAN learn the target distribution from noise input results in a better model generalization during the test time and allows the model to fit almost perfectly to the target domain distribution. As a result, our model surpasses the Pix2Pix model in segmenting HC18 and Montgomery's chest x-ray images. Both qualitative and Dice scores show the superiority of our model. Although our proposed method does not use thousand of additional data for pretraining, it produces comparable results for the in and out-domain generalization compared to the state-of-the-art methods.

cs.CV↗

arXiv2022

OAIR: Object-Aware Image Retargeting Using PSO and Aesthetic Quality Assessment

Image retargeting aims at altering an image size while preserving important content and minimizing noticeable distortions. However, previous image retargeting methods create outputs that suffer from artifacts and distortions. Besides, most previous works attempt to retarget the background and foreground of the input image simultaneously. Simultaneous resizing of the foreground and background causes changes in the aspect ratios of the objects. The change in the aspect ratio is specifically not desirable for human objects. We propose a retargeting method that overcomes these problems. The proposed approach consists of the following steps. Firstly, an inpainting method uses the input image and the binary mask of foreground objects to produce a background image without any foreground objects. Secondly, the seam carving method resizes the background image to the target size. Then, a super-resolution method increases the input image quality, and we then extract the foreground objects. Finally, the retargeted background and the extracted super-resolued objects are fed into a particle swarm optimization algorithm (PSO). The PSO algorithm uses aesthetic quality assessment as its objective function to identify the best location and size for the objects to be placed in the background. We used image quality assessment and aesthetic quality assessment measures to show our superior results compared to popular image retargeting techniques.

cs.CV↗

arXiv2022

Adaptive search area for fast motion estimation

This paper suggests a new method for determining the search area for a motion estimation algorithm based on block matching. The search area is adaptively found in the proposed method for each frame block. This search area is similar to that of the full search (FS) algorithm but smaller for most blocks of a frame. Therefore, the proposed algorithm is analogous to FS in terms of regularity but has much less computational complexity. The temporal and spatial correlations among the motion vectors of blocks are used to find the search area. The matched block is chosen from a rectangular area that the prediction vectors set out. Simulation results indicate that the speed of the proposed algorithm is at least seven times better than the FS algorithm.

cs.CV↗

arXiv2021

MSGDD-cGAN: Multi-Scale Gradients Dual Discriminator Conditional Generative Adversarial Network

Conditional Generative Adversarial Networks (cGANs) have been used in many image processing tasks. However, they still have serious problems maintaining the balance between conditioning the output on the input and creating the output with the desired distribution based on the corresponding ground truth. The traditional cGANs, similar to most conventional GANs, suffer from vanishing gradients, which backpropagate from the discriminator to the generator. Moreover, the traditional cGANs are sensitive to architectural changes due to previously mentioned gradient problems. Therefore, balancing the architecture of the cGANs is almost impossible. Recently MSG-GAN has been proposed to stabilize the performance of the GANs by applying multiple connections between the generator and discriminator. In this work, we propose a method called MSGDD-cGAN, which first stabilizes the performance of the cGANs using multi-connections gradients flow. Secondly, the proposed network architecture balances the correlation of the output to input and the fitness of the output on the target distribution. This balance is generated by using the proposed dual discrimination procedure. We tested our model by segmentation of fetal ultrasound images. Our model shows a 3.18% increase in the F1 score comparing to the pix2pix version of cGANs.

cs.CV↗

arXiv2021

Robust Watermarking using Diffusion of Logo into Autoencoder Feature Maps

Digital contents have grown dramatically in recent years, leading to increased attention to copyright. Image watermarking has been considered one of the most popular methods for copyright protection. With the recent advancements in applying deep neural networks in image processing, these networks have also been used in image watermarking. Robustness and imperceptibility are two challenging features of watermarking methods that the trade-off between them should be satisfied. In this paper, we propose to use an end-to-end network for watermarking. We use a convolutional neural network (CNN) to control the embedding strength based on the image content. Dynamic embedding helps the network to have the lowest effect on the visual quality of the watermarked image. Different image processing attacks are simulated as a network layer to improve the robustness of the model. Our method is a blind watermarking approach that replicates the watermark string to create a matrix of the same size as the input image. Instead of diffusing the watermark data into the input image, we inject the data into the feature space and force the network to do this in regions that increase the robustness against various attacks. Experimental results show the superiority of the proposed method in terms of imperceptibility and robustness compared to the state-of-the-art algorithms.

cs.MM↗

arXiv2020

Literature Review of Computer Tools for the Visually Impaired: a focus on Search Engines

A sudden reliance on the internet has resulted in the global standardization of specific software and interfaces tailored for the average user. Whether it be web apps or dedicated software, the methods of interaction are seemingly similar. But when the computer tool is presented with unique users, specifically with a disability, the quality of interaction degrades, sometimes to a point of complete uselessness. This roots from one's focus on the average user rather than the development of a platform for all (a golden standard). This paper reviews published works and products that deal with providing accessibility to visually impaired online users. Due to the variety of tools that are available to computer users, the paper focuses on search engines as a primary tool for browsing the web. By analyzing the attributes discussed below, the reader is equipped with a set of references for existing applications, along with practical insight and recommendations for accessible design. Finally, the necessary considerations for future developments and summaries of important focal points are highlighted.

cs.HC↗

arXiv2020

Classification of Diabetic Retinopathy Using Unlabeled Data and Knowledge Distillation

Knowledge distillation allows transferring knowledge from a pre-trained model to another. However, it suffers from limitations, and constraints related to the two models need to be architecturally similar. Knowledge distillation addresses some of the shortcomings associated with transfer learning by generalizing a complex model to a lighter model. However, some parts of the knowledge may not be distilled by knowledge distillation sufficiently. In this paper, a novel knowledge distillation approach using transfer learning is proposed. The proposed method transfers the entire knowledge of a model to a new smaller one. To accomplish this, unlabeled data are used in an unsupervised manner to transfer the maximum amount of knowledge to the new slimmer model. The proposed method can be beneficial in medical image analysis, where labeled data are typically scarce. The proposed approach is evaluated in the context of classification of images for diagnosing Diabetic Retinopathy on two publicly available datasets, including Messidor and EyePACS. Simulation results demonstrate that the approach is effective in transferring knowledge from a complex model to a lighter one. Furthermore, experimental results illustrate that the performance of different small models is improved significantly using unlabeled data and knowledge distillation.

eess.IV↗

arXiv2020

Hardware Implementation of Adaptive Watermarking Based on Local Spatial Disorder Analysis

With the increasing use of the internet and the ease of exchange of multimedia content, the protection of ownership rights has become a significant concern. Watermarking is an efficient means for this purpose. In many applications, real-time watermarking is required, which demands hardware implementation of low complexity and robust algorithm. In this paper, an adaptive watermarking is presented, which uses embedding in different bit-planes to achieve transparency and robustness. Local disorder of pixels is analyzed to control the strength of the watermark. A new low complexity method for disorder analysis is proposed, and its hardware implantation is presented. An embedding method is proposed, which causes lower degradation in the watermarked image. Also, the performance of proposed watermarking architecture is improved by a pipe-line structure and is tested on an FPGA device. Results show that the algorithm produces transparent and robust watermarked images. The synthesis report from FPGA implementation illustrates a low complexity hardware structure.

cs.MM↗

arXiv2020

A fast semi-automatic method for classification and counting the number and types of blood cells in an image

A novel and fast semi-automatic method for segmentation, locating and counting blood cells in an image is proposed. In this method, thresholding is used to separate the nucleus from the other parts. We also use Hough transform for circles to locate the center of white cells. Locating and counting of red cells is performed using template matching. We make use of finding local maxima, labeling and mean value computation in order to shrink the areas obtained after applying Hough transform or template matching, to a single pixel as representative of location of each region. The proposed method is very fast and computes the number and location of white cells accurately. It is also capable of locating and counting the red cells with a small error.

eess.IV↗

arXiv2020

Acceleration of Convolutional Neural Network Using FFT-Based Split Convolutions

Convolutional neural networks (CNNs) have a large number of variables and hence suffer from a complexity problem for their implementation. Different methods and techniques have developed to alleviate the problem of CNN's complexity, such as quantization, pruning, etc. Among the different simplification methods, computation in the Fourier domain is regarded as a new paradigm for the acceleration of CNNs. Recent studies on Fast Fourier Transform (FFT) based CNN aiming at simplifying the computations required for FFT. However, there is a lot of space for working on the reduction of the computational complexity of FFT. In this paper, a new method for CNN processing in the FFT domain is proposed, which is based on input splitting. There are problems in the computation of FFT using small kernels in situations such as CNN. Splitting can be considered as an effective solution for such issues aroused by small kernels. Using splitting redundancy, such as overlap-and-add, is reduced and, efficiency is increased. Hardware implementation of the proposed FFT method, as well as different analyses of the complexity, are performed to demonstrate the proper performance of the proposed method.

cs.CV↗

arXiv2020

Convolutional Neural Network Pruning Using Filter Attenuation

Filters are the essential elements in convolutional neural networks (CNNs). Filters are corresponded to the feature maps and form the main part of the computational and memory requirement for the CNN processing. In filter pruning methods, a filter with all of its components, including channels and connections, are removed. The removal of a filter can cause a drastic change in the network's performance. Also, the removed filters cannot come back to the network structure. We want to address these problems in this paper. We propose a CNN pruning method based on filter attenuation in which weak filters are not directly removed. Instead, weak filters are attenuated and gradually removed. In the proposed attenuation approach, weak filters are not abruptly removed, and there is a chance for these filters to return to the network. The filter attenuation method is assessed using the VGG model for the Cifar10 image classification task. Simulation results show that the filter attenuation works with different pruning criteria, and better results are obtained in comparison with the conventional pruning methods.

cs.CV↗

arXiv2020

Splitting Convolutional Neural Network Structures for Efficient Inference

For convolutional neural networks (CNNs) that have a large volume of input data, memory management becomes a major concern. Memory cost reduction can be an effective way to deal with these problems that can be realized through different techniques such as feature map pruning, input data splitting, etc. Among various methods existing in this area of research, splitting the network structure is an interesting research field, and there are a few works done in this area. In this study, the problem of reducing memory utilization using network structure splitting is addressed. A new technique is proposed to split the network structure into small parts that consume lower memory than the original network. The split parts can be processed almost separately, which provides an essential role for better memory management. The split approach has been tested on two well-known network structures of VGG16 and ResNet18 for the classification of CIFAR10 images. Simulation results show that the splitting method reduces both the number of computational operations as well as the amount of memory consumption.

cs.CV↗