Hossein Rahmani

• PhD
• Senior Lecturer at Lancaster University

Depth sensors open up possibilities of dealing with the human action recognition problem by providing 3D human skeleton data and depth images of the scene. Analysis of human actions based on 3D skeleton data has become popular recently, due to its robustness and view-invariant representation. However, the skeleton alone is insufficient to distingui...

Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which are viewpoint dependent. In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views. We propose the Histogram of Oriented Principal Components (HOPC) descriptor t...

Recognizing human actions from unknown and unseen (novel) views is a challenging problem. We propose a Robust Non-Linear Knowledge Transfer Model (R-NKTM) for human action recognition from novel views. The proposed R-NKTM is a deep fully-connected neural network that transfers knowledge of human actions from any unknown view to a shared high-level...

AI-generated content (AIGC) methods aim to produce text, images, videos, 3D assets, and other media using AI algorithms. Due to its wide range of applications and the potential of recent works, AIGC developments – especially in Machine Learning (ML) and Deep Learning (DL) – have been attracting significant attention, and this survey focuses on comp...

Video diffusion models have recently achieved remarkable results in video generation. Despite their encouraging performance, most of these models are mainly designed and trained for short video generation, leading to challenges in maintaining temporal consistency and visual details in long video generation. In this paper, we propose LongDiff, a nov...

Human-object interaction (HOI) detection often faces high levels of ambiguity and indeterminacy, as the same interaction can appear vastly different across different human-object pairs. Additionally, the indeterminacy can be further exacerbated by issues such as occlusions and cluttered backgrounds. To handle such a challenging task, in this work,...

In contrast to batch learning where all training data is available at once, continual learning represents a family of methods that accumulate knowledge and learn continuously with data available in sequential order. Similar to the human learning process with the ability of learning, fusing and accumulating new knowledge acquired at different time s...

Images of the human hand can be effectively deployed to assist with the identification of the perpetrators of serious crimes. One of the prominent and distinguishing features of the human hand is found in the skin of the finger knuckle regions, which includes creases forming complex and distinctive patterns. Exploiting knuckle skin crease patternin...

Text-based editing of 3D human avatars to precisely match user requirements is challenging due to the inherent ambiguity and limited expressiveness of natural language. To overcome this, we propose the Avatar Concept Slider (ACS), a 3D avatar editing method that allows precise editing of semantic concepts in human avatars towards a specified interm...

Recently, Gaussian Splatting, a method that represents a 3D scene as a collection of Gaussian distributions, has gained significant attention in addressing the task of novel view synthesis. In this paper, we highlight a fundamental limitation of Gaussian Splatting: its inability to accurately render discontinuities and boundaries in images due to t...

Creating and customizing a 3D clothed avatar from textual descriptions is a critical and challenging task. Traditional methods often treat the human body and clothing as inseparable, limiting users' ability to freely mix and match garments. In response to this limitation, we present LAyered Gaussian Avatar (LAGA), a carefully designed framework ena...

AI-generated content (AIGC) methods aim to produce text, images, videos, 3D assets, and other media using AI algorithms. Due to its wide range of applications and the potential of recent works, AIGC developments -- especially in Machine Learning (ML) and Deep Learning (DL) -- have been attracting significant attention, and this survey focuses on co...

Recovering a 3D human mesh from a single RGB image is a challenging task due to depth ambiguity and self-occlusion, resulting in a high degree of uncertainty. Meanwhile, diffusion models have recently seen much success in generating high-quality outputs by progressively denoising noisy inputs. Inspired by their capability, we explore a diffusion-ba...

Recently, deep learning based approaches have shown promising results in 3D hand reconstruction from a single RGB image. These approaches can be roughly divided into model-based approaches, which are heavily dependent on the model's parameter space, and model-free approaches, which require large numbers of 3D ground truths to reduce depth ambiguity...

Learning with large-scale unlabeled data has become a powerful tool for pre-training Visual Transformers (VTs). However, prior works tend to overlook that, in real-world scenarios, the input data may be corrupted and unreliable. Pre-training VTs on such corrupted data can be challenging, especially when we pre-train via the masked autoencoding appr...

Currently, salience-based channel pruning makes continuous breakthroughs in network compression. In the realization, the salience mechanism is used as a metric of channel salience to guide pruning. Therefore, salience-based channel pruning can dynamically adjust the channel width at run-time, which provides a flexible pruning scheme. However, there...

Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against energy-oriented attacks targeted at reducing their efficiency. Specifically, we attack dynamic models with our novel algorithm Gr...

Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against energy-oriented attacks targeted at reducing their efficiency. Specifically, we attack dynamic models with our novel algorithm...

Currently, salience-based channel pruning makes continuous breakthroughs in network compression. In the realization, the salience mechanism is used as a metric of channel salience to guide pruning. Therefore, salience-based channel pruning can dynamically adjust the channel width at run-time, which provides a flexible pruning scheme. However, there...

Monocular 3D human pose estimation is quite challenging due to the inherent ambiguity and occlusion, which often lead to high uncertainty and indeterminacy. On the other hand, diffusion models have recently emerged as an effective tool for generating high-quality images from noise. Inspired by their capability, we explore a novel pose estimation fr...

Despite extensive research, 3D face reconstruction from a single image remains an open research problem due to the high degree of variability in pose, occlusions and complex lighting conditions. While deep learning-based methods have achieved great success, they are usually limited to near frontal images and images that are free of occlusions. Also...

Early action prediction aims to successfully predict the class label of an action before it is completely performed. This is a challenging task because the beginning stages of different actions can be very similar, with only minor subtle differences for discrimination. In this paper, we propose a novel Expert Retrieval and Assembly (ERA) module tha...

Human interaction recognition is very important in many applications. One crucial cue in recognizing an interaction is the interactive body parts. In this work, we propose a novel Interaction Graph Transformer (IGFormer) network for skeleton-based interaction recognition via modeling the interactive body parts as graphs. More specifically, the prop...

The goal of fine-grained action recognition is to successfully discriminate between action categories with subtle differences. To tackle this, we derive inspiration from the human visual system which contains specialized regions in the brain that are dedicated towards handling specific tasks. We design a novel Dynamic Spatio-Temporal Specialization...

Dynamic neural networks could adapt their structures or parameters based on different inputs. By reducing the computation redundancy for certain samples, it can greatly improve the computational efficiency without compromising the accuracy. In this paper, we investigate the robustness of dynamic neural networks against energy-oriented attacks. We p...

The goal of fine-grained action recognition is to successfully discriminate between action categories with subtle differences. To tackle this, we derive inspiration from the human visual system which contains specialized regions in the brain that are dedicated towards handling specific tasks. We design a novel Dynamic Spatio-Temporal Specialization...

In this paper, we propose a novel hand-based person recognition method for the purpose of criminal investigations since the hand image is often the only available information in cases of serious crime such as sexual abuse. Our proposed method, Multi-Branch with Attention Network (MBA-Net), incorporates both channel and spatial attention modules in...

Human interaction recognition is very important in many applications. One crucial cue in recognizing an interaction is the interactive body parts. In this work, we propose a novel Interaction Graph Transformer (IGFormer) network for skeleton-based interaction recognition via modeling the interactive body parts as graphs. More specifically, the prop...

Early action prediction aims to successfully predict the class label of an action before it is completely performed. This is a challenging task because the beginning stages of different actions can be very similar, with only minor subtle differences for discrimination. In this paper, we propose a novel Expert Retrieval and Assembly (ERA) module tha...

Existing approaches for 2D pose estimation in videos often require a large number of dense annotations, which are costly and labor intensive to acquire. In this paper, we propose a semi-supervised REinforced MOtion Transformation nEtwork (REMOTE) to leverage a few labeled frames and temporal pose variations in videos, which enables effective learni...

Human Action Recognition (HAR) aims to understand human behavior and assign a label to each action. It has a wide range of applications, and therefore has been attracting increasing attention in the field of computer vision. Human actions can be represented using various data modalities, such as RGB, skeleton, depth, infrared, point cloud, event st...

In cases of serious crime, including sexual abuse, often the only available information with demonstrated potential for identification is images of the hands. Since this evidence is captured in uncontrolled situations, it is difficult to analyse. As global approaches to feature comparison are limited in this case, it is important to extend to consi...

This poster was used to depict our work on knuckle localization and identification, carried out under the flagship of H-Unique Project.

The rapid outbreak of coronavirus threatens humans’ life all around the world. Due to the insufficient diagnostic infrastructures, developing an accurate, efficient, inexpensive, and quick diagnostic tool is of great importance. To date, researchers have proposed several detection models based on chest imaging analysis, primarily based on deep neur...

The knuckle creases present on the dorsal side of the human hand can play significant role in identifying the offenders of serious crime, especially when evidence images of more recognizable biometric traits, such as the face, are not available. These knuckle creases, if localized appropriately, can result in improved identification ability. This i...

Most of the state-of-the-art action recognition methods focus on offline learning, where the samples of all types of actions need to be provided at once. Here, we address continual learning of action recognition, where various types of new actions are continuously learned over time.This task is quite challenging, owing to the catastrophic forgettin...

In contrast to batch learning where all training data is available at once, continual learning represents a family of methods that accumulate knowledge and learn continuously with data available in sequential order. Similar to the human learning process with the ability of learning, fusing, and accumulating new knowledge coming at different time st...

In contrast to batch learning where all training data is available at once, continual learning represents a family of methods that accumulate knowledge and learn continuously with data available in sequential order. Similar to the human learning process with the ability of learning, fusing, and accumulating new knowledge coming at different time st...

In this paper, we introduce the Multi-Modal Video Reasoning and Analyzing Competition (MMVRAC) workshop in conjunction with ICCV 2021. This competition is composed of four different tracks, namely, video question answering , skeleton-based action recognition, fisheye video-based action recognition, and person re-identification, which are based on t...

In this paper, we introduce the Multi-Modal Video Reasoning and Analyzing Competition (MMVRAC) workshop in conjunction with ICCV 2021. This competition is composed of four different tracks, namely, video question answering, skeleton-based action recognition, fisheye video-based action recognition, and person re-identification, which are based on tw...

In this paper, we propose a novel hand-based person recognition method for the purpose of criminal investigations since the hand image is often the only available information in cases of serious crime such as sexual abuse. Our proposed method, Multi-Branch with Attention Network (MBA-Net), incorporates both channel and spatial attention modules in...

In many cases of serious crime, images of a hand can be the only evidence available for the forensic identification of the offender. As well as placing them at the scene, such images and video evidence offer proof of the offender committing the crime. The knuckle creases of the human hand have emerged as an effective biometric trait and been used t...

Hyperspectral image (HSI) classification is an important topic in the community of remote sensing, which has a wide range of applications in geoscience. Recently, deep learning-based methods have been widely used in HSI classification. However, due to the scarcity of labeled samples in HSI, the potential of deep learning-based methods has not been...

This is a survey on human action recognition from various data modalities.

In cases of serious crime, including sexual abuse, often the only available information with demonstrated potential for identification is images of the hands. Since this evidence is captured in uncontrolled situations, it is difficult to analyse. As global approaches to feature comparison are limited in this case, it is important to extend to consi...

The rapid outbreak of COVID-19 threatens humans life all around the world. Due to insufficient diagnostic infrastructures, developing an accurate, efficient, inexpensive, and quick diagnostic tool is of great importance. As chest radiography, such as chest X-ray (CXR) and CT computed tomography (CT), is a possible way for screening COVID-19, develo...

Human Action Recognition (HAR), aiming to understand human behaviors and then assign category labels, has a wide range of applications, and thus has been attracting increasing attention in the field of computer vision. Generally, human actions can be represented using various data modalities, such as RGB, skeleton, depth, infrared sequence, point c...

Regularization plays a vital role in the context of deep learning by preventing deep neural networks from the danger of overfitting. This paper proposes a novel deep learning regularization method named as DL-Reg, which carefully reduces the nonlinearity of deep networks to a certain extent by explicitly enforcing the network to behave as much line...

Audio, animations and video belong to a class of data known as delay sensitive because they are sensitive to delays in presentation to the users. Also, because of huge data in such items, disk is an important device in managing them. In order to have an acceptable presentation, disk requests deadlines must be met, and a real-time scheduling approac...

We propose Human Pose Models that represent RGB and depth images of human poses independent of clothing textures, backgrounds, lighting conditions , body shapes and camera viewpoints. Learning such universal models requires training images where all factors are varied for every human pose. Capturing such data is prohibitively expensive. Therefore,...

Human actions represented with 3D skeleton sequences are robust to clustered backgrounds and illumination changes. In this paper, we investigate skeleton-based action prediction, which aims to recognize an action from a partial skeleton sequence that contains incomplete action information. We propose a new Latent Global Network based on adversarial...

The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched imag...

We propose an image-based technique to measure the volume, weight and the size distribution of gravel particles in a gravel-soil mixture. The proposed method uses 3D scanning and a surface reconstruction algorithm to generate a high-resolution depth image, which is then used to accurately estimate the volume and weight of each gravel particle. The...

Automation is becoming a large component of many industries in the 21st century, in areas ranging from manufacturing, communications and transportation. Automation has offered promised returns of improvements in safety, productivity and reduced costs. Many industry leaders are specifically working on the application of autonomous technology in tran...

In this paper, we propose a new approach to recognize the class label of an action before this action is fully performed based on skeleton sequences. Compared to action recognition which uses fully observed action sequences, early action recognition with partial sequences is much more challenging mainly due to: (1) the global information of a long-...

Person re-identification (re-ID), which aims to identify the same individual from a gallery collected with different cameras, has attracted increasing attention in the multimedia retrieval community. Current deep learning methods for person re-identification (re-ID) focus on learning classification models on training identities to obtain a ID-discr...

The measurement of bulk density in gravelly soils (>15% soil particles >2 mm) is more time‐consuming than for other soils. The excavation method, usually employed for measurement of bulk density in gravelly soils, includes excavating a void and calculating volume of the void from the weight and density of the material (e.g. sand and plaster cast) u...

We discussed neural networks in Chapter 3. CNNs are one of the most popular categories of neural networks, especially for high-dimensional data (e.g., images and videos). CNNs operate in a way that is very similar to standard neural networks. A key difference, however, is that each unit in a CNN layer is a two- (or high-) dimensional filter which i...

We have covered the basic modules in the previous chapters which can be joined together to develop CNN-based deep learning models. Among these modules, we covered convolution, subsampling and several other layers which form large-scale CNN architectures. We noticed that the loss functions are used during training to measure the difference between t...

There have been a lot of interest from academics (e.g., The University of California Berkeley, New York University, The University of Toronto, The University of Montreal) and industry groups (e.g., Google, Facebook, Microsoft) to develop deep learning frameworks. It is mainly due to their popularity in many applications domains over the last few ye...

The application of deep learning algorithms, especially CNNs, to computer vision problems have seen a rapid progress. This has led to highly robust, efficient, and flexible vision systems. This book aimed to introduce different aspects of CNNs in computer vision problems. The first part of this book (Chapter 1 and Chapter 2) introduced computer vis...

In Chapter 4, we discussed different architecture blocks of the CNN and their operational details. Most of these CNN layers involve parameters which are required to be tuned appropriately for a given computer vision task (e.g., image classification and object detection). In this chapter, we will discuss various mechanisms and techniques that are us...

Before going into the details of the CNNs, we provide in this chapter an introduction to artificial neural networks, their computational mechanism, and their historical background. Neural networks are inspired by the working of cerebral cortex in mammals. It is important to note, however, that these models do not closely resemble the working, scale...

Feature extraction and classification are two key stages of a typical computer vision system. In this chapter, we provide an introduction to these two steps: their importance and their design challenges for computer vision tasks.

Computer vision is a very broad research area which covers a wide variety of approaches not only to process images but also to understand their contents. It is an active research field for convolutional neural network applications. The most popular of these applications include, classification, segmentation, detection, and scene understanding. Most...

Recognizing human actions from unknown and unseen (novel) views is a challenging problem. We propose a Robust Non-Linear Knowledge Transfer Model (R-NKTM) for human action recognition from novel views. The proposed R-NKTM is a deep fully-connected neural network that transfers knowledge of human actions from any unknown view to a shared high-level...

This paper concerns action recognition from unseen and unknown views. We propose unsupervised learning of a non-linear model that transfers knowledge from multiple views to a canonical view. The proposed Non-linear Knowledge Transfer Model (NKTM) is a deep network, with weight decay and sparsity constraints, which finds a shared high-level virtual...

Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which change significantly with viewpoint. In contrast, we directly process the pointclouds and propose a new technique for action recognition which is more robust to noise, action speed and viewpoint variations. Our t...

We propose an action classification algorithm which uses Locality-constrained Linear Coding (LLC) to capture discriminative information of human body variations in each spatiotemporal subsequence of a video sequence. Our proposed method divides the input video into equally spaced overlapping spatiotemporal subsequences, each of which is decomposed...

We propose an algorithm which combines the discriminative information from depth images as well as from 3D joint positions to achieve high action recognition accuracy. To avoid the suppression of subtle discriminative information and also to handle local occlusions, we compute a vector of many independent local features. Each feature encodes spatio...

Harmful contents are rising in internet day by day and this motivates the essence of more research in fast and reliable obscene and immoral material filtering. Pornographic image recognition is an important component in each filtering system. In this paper, a new approach for detecting pornographic images is introduced. In this approach, two new fe...

Disk management is an increasingly important aspect of operating systems research and development because it has great effect on system performance. As the gap between processor and disk performance continues to increase in modern systems, access to mass storage is a common bottleneck that ultimately limits overall system performance. In this paper...

In digital image watermarking, an image is embedded into a picture for a variety of purposes such as captioning and copyright protection. In this paper, a robust private watermarking scheme for embedding a gray-scale watermark is proposed. In the proposed method, the watermark and original image are processed by applying blockwise DCT. Also, a Dyna...

Disk scheduling has an important role in QOS guarantee of soft real-time environments such as video-on-demand and multimedia servers. Since now, some disk-scheduling algorithms have been proposed to schedule real-time disk requests. One of the most recent algorithms is global seek-optimizing real-time (GSR) that schedules the disk requests with dif...

Disk scheduling is an operating system process to service disk requests. It has an important role in QOS guarantee of soft real-time environments such as video-on-demand and multimedia servers. Since now, some disk scheduling algorithms have been proposed to schedule disk requests in an optimized manner. Most of these methods try to minimize makesp...

This paper presents a new semi-blind robust image watermarking scheme based on most perceptually important sub-image (MPISI) of watermark and original image. The proposed method is based on lossless approach and to prevent false-positive watermark detection in this approach, a watermark certificate is inserted in the image. Therefore, the proposed...

Watermarking techniques have been proposed to solve the problem of copyright protection and authentication of digital media. In this paper, a robust blind lossless watermarking scheme is proposed that is based on similarity between DCT coefficients of original image and watermark. The watermark extraction procedure does not require the original ima...

Disk scheduling has an important role in QOS guarantee of soft real-time environments such as video-on-demand and multimedia servers. Since now, some disk scheduling algorithms have been proposed to optimize scheduling disk requests. One of the most recent algorithms is GSR. GSR improved the disk throughput by globally rescheduling scheme for real-...