Thangarajah Akilan

Kanchan Keisham, Thenukan Pathmanathan, Thangarajah Akilan

Zero-shot Temporal Action Localization (ZS-TAL) aims to detect and locate previously unseen actions in untrimmed videos. However, existing approaches primarily focus on modeling long-range contextual information, often neglecting the critical relative-offset-based local correlations between video frames. Furthermore, their performance is hindered by limited feature representation capabilities due to the shallow nature of their network architectures. In this paper, we address these limitations by introducing a novel local-global multi-scale feature representation module. We propose a novel multi-scale encoder architecture, termed ConTrans, that integrates convolutional (Conv) inductive biases with transformer Self-attention to jointly capture fine-grained local dependencies and long-range global context, leading to more comprehensive feature representations than existing methods. Experimental evaluations on the ActivityNet-1.3 and THUMOS14 datasets demonstrate that ConTrans significantly outperforms existing methods, establishing a new benchmark for ZS-TAL.

Thenukan Pathmanathan, Kanchan Keisham, Thangarajah Akilan

Point cloud segmentation is a fundamental task in 3D scene understanding. Its progress is constrained by the high cost and time required for dense 3D annotations, making labeled samples difficult to obtain. Beyond annotation scarcity, different sensing modalities face inherent limitations. 2D images provide rich texture and appearance cues, yet they lack explicit depth and geometric structure. In contrast, 3D point clouds capture accurate spatial geometry but are sparse and contain no texture information. As a result, relying on a single modality restricts the richness of learned representations and weakens generalization. Although recent multi-modal methods that combine 3D point clouds with 2D images have demonstrated strong performance in tasks such as classification and retrieval, they typically depend on large-scale labeled datasets and have not been fully exploited for data-efficient dense prediction. To address these limitations, we propose a novel cross-modal knowledge distillation framework, xModel-KD, for 3D point cloud segmentation. Our method exploits the complementary strengths of 2D texture and 3D geometry by learning unified per-point representations through cross-modal alignment. Specifically, we design a cross-modal fusion encoder trained with a contrastive objective that enforces feature consistency between corresponding 2D and 3D representations across multiple views. By integrating powerful pre-trained backbones with a targeted fusion strategy, the proposed framework effectively transfers appearance cues from images to geometry-aware point features. Experimental results show that cross-modal fusion achieves a 2% absolute improvement in mIoU over a LiDAR-only baseline, demonstrating the benefit of leveraging complementary multi-modal information for scalable and annotation-efficient 3D scene understanding.

Dhrumil Patel, Dhruv Patel, Rudra Saxena et al.

Brain tumor segmentation from magnetic resonance imaging (MRI) plays an important role in diagnostic radiology. To overcome the practical issues in manual approaches, there is a huge demand for building automatic tumor segmentation algorithms. This work introduces an efficient brain tumor summation model by exploiting the advancement in MRI and graph neural networks (GNNs). The model represents the volumetric MRI as a region adjacency graph (RAG) and learns to identify the type of tumors through a graph attention network (GAT) -- a variant of GNNs. The ablation analysis conducted on two benchmark datasets proves that the proposed model can produce competitive results compared to the leading-edge solutions. It achieves mean dice scores of 0.91, 0.86, 0.79, and mean Hausdorff distances in the 95th percentile (HD95) of 5.91, 6.08, and 9.52 mm, respectively, for whole tumor, core tumor, and enhancing tumor segmentation on BraTS2021 validation dataset. On average, these performances are >6\% and >50%, compared to a GNN-based baseline model, respectively, on dice score and HD95 evaluation metrics.

Jun Li, Thangarajah Akilan

Temperature monitoring is critical for electrical motors to determine if device protection measures should be executed. However, the complexity of the internal structure of Permanent Magnet Synchronous Motors (PMSM) makes the direct temperature measurement of the internal components difficult. This work pragmatically develops three deep learning models to estimate the PMSMs' internal temperature based on readily measurable external quantities. The proposed supervised learning models exploit Long Short-Term Memory (LSTM) modules, bidirectional LSTM, and attention mechanism to form encoder-decoder structures to predict simultaneously the temperatures of the stator winding, tooth, yoke, and permanent magnet. Experiments were conducted in an exhaustive manner on a benchmark dataset to verify the proposed models' performances. The comparative analysis shows that the proposed global attention-based encoder-decoder (EnDec) model provides a competitive overall performance of 1.72 Mean Squared Error (MSE) and 5.34 Mean Absolute Error (MAE).

Siavash Hosseini, Thangarajah Akilan

Global oil demand is rapidly increasing and is expected to reach 106.3 million barrels per day by 2040. Thus, it is vital for hydrocarbon extraction industries to forecast their production to optimize their operations and avoid losses. Big companies have realized that exploiting the power of deep learning (DL) and the massive amount of data from various oil wells for this purpose can save a lot of operational costs and reduce unwanted environmental impacts. In this direction, researchers have proposed models using conventional machine learning (ML) techniques for oil production forecasting. However, these techniques are inappropriate for this problem as they can not capture historical patterns found in time series data, resulting in inaccurate predictions. This research aims to overcome these issues by developing advanced data-driven regression models using sequential convolutions and long short-term memory (LSTM) units. Exhaustive analyses are conducted to select the optimal sequence length, model hyperparameters, and cross-well dataset formation to build highly generalized robust models. A comprehensive experimental study on Volve oilfield data validates the proposed models. It reveals that the LSTM-based sequence learning model can predict oil production better than the 1-D convolutional neural network (CNN) with mean absolute error (MAE) and R2 score of 111.16 and 0.98, respectively. It is also found that the LSTM-based model performs better than all the existing state-of-the-art solutions and achieves a 37% improvement compared to a standard linear regression, which is considered the baseline model in this work.

Thangarajah Akilan, Hrishikesh Vachhani

Pedestrian detection has become a cornerstone for several high-level tasks, including autonomous driving, intelligent transportation, and traffic surveillance. There are several works focussed on pedestrian detection using visible images, mainly in the daytime. However, this task is very intriguing when the environmental conditions change to poor lighting or nighttime. Recently, new ideas have been spurred to use alternative sources, such as Far InfraRed (FIR) temperature sensor feeds for detecting pedestrians in low-light conditions. This study reviews recent developments in low-light pedestrian detection approaches. It systematically categorizes and analyses various algorithms from region-based to non-region-based and graph-based learning methodologies by highlighting their methodologies, implementation issues, and challenges. It also outlines the key benchmark datasets that can be used for research and development of advanced pedestrian detection algorithms, particularly in low-light situations.

Mohammed Nowshad Ruhani Chowdhury, Wandong Zhang, Thangarajah Akilan

Studies show that Studies that cardiovascular diseases (CVDs) are malignant for human health. Thus, it is important to have an efficient way of CVD prognosis. In response to this, the healthcare industry has adopted machine learning-based smart solutions to alleviate the manual process of CVD prognosis. Thus, this work proposes an information fusion technique that combines key attributes of a person through analysis of variance (ANOVA) and domain experts' knowledge. It also introduces a new collection of CVD data samples for emerging research. There are thirty-eight experiments conducted exhaustively to verify the performance of the proposed framework on four publicly available benchmark datasets and the newly created dataset in this work. The ablation study shows that the proposed approach can achieve a competitive mean average accuracy (mAA) of 99.2% and a mean average AUC of 97.9%.

Mei Ling Chee, Thangarajah Akilan, Aparna Ravindra Phalke et al.

Semantic segmentation in high-resolution agricultural imagery demands models that strike a careful balance between accuracy and computational efficiency to enable deployment in practical systems. In this work, we propose DAS-SK, a novel lightweight architecture that retrofits selective kernel convolution (SK-Conv) into the dual atrous separable convolution (DAS-Conv) module to strengthen multi-scale feature learning. The model further enhances the atrous spatial pyramid pooling (ASPP) module, enabling the capture of fine-grained local structures alongside global contextual information. Built upon a modified DeepLabV3 framework with two complementary backbones - MobileNetV3-Large and EfficientNet-B3, the DAS-SK model mitigates limitations associated with large dataset requirements, limited spectral generalization, and the high computational cost that typically restricts deployment on UAVs and other edge devices. Comprehensive experiments across three benchmarks: LandCover.ai, VDD, and PhenoBench, demonstrate that DAS-SK consistently achieves state-of-the-art performance, while being more efficient than CNN-, transformer-, and hybrid-based competitors. Notably, DAS-SK requires up to 21x fewer parameters and 19x fewer GFLOPs than top-performing transformer models. These findings establish DAS-SK as a robust, efficient, and scalable solution for real-time agricultural robotics and high-resolution remote sensing, with strong potential for broader deployment in other vision domains.

Thangarajah Akilan, Nusrat Jahan, Wandong Zhang

Supervised learning demands large amounts of precisely annotated data to achieve promising results. Such data curation is labor-intensive and imposes significant overhead regarding time and costs. Self-supervised learning (SSL) partially overcomes these limitations by exploiting vast amounts of unlabeled data and creating surrogate (pretext or proxy) tasks to learn useful representations without manual labeling. As a result, SSL has become a powerful machine learning (ML) paradigm for solving several practical downstream computer vision problems, such as classification, detection, and segmentation. Image segmentation is the cornerstone of many high-level visual perception applications, including medical imaging, intelligent transportation, agriculture, and surveillance. Although there is substantial research potential for developing advanced algorithms for SSL-based semantic segmentation, a comprehensive study of existing methodologies is essential to trace advances and guide emerging researchers. This survey thoroughly investigates over 150 recent image segmentation articles, particularly focusing on SSL. It provides a practical categorization of pretext tasks, downstream tasks, and commonly used benchmark datasets for image segmentation research. It concludes with key observations distilled from a large body of literature and offers future directions to make this research field more accessible and comprehensible for readers.

Christopher Silver, Thangarajah Akilan

Falls among seniors are a major public health issue. Existing solutions using wearable sensors, ambient sensors, and RGB-based vision systems face challenges in reliability, user compliance, and practicality. Studies indicate that stakeholders, such as older adults and eldercare facilities, prefer non-wearable, passive, privacy-preserving, and real-time fall detection systems that require no user interaction. This study proposes an advanced thermal fall detection method using a Bidirectional Convolutional Long Short-Term Memory (BiConvLSTM) model, enhanced with spatial, temporal, feature, self, and general attention mechanisms. Through systematic experimentation across hundreds of model variations exploring the integration of attention mechanisms, recurrent modules, and motion flow, we identified top-performing architectures. Among them, BiConvLSTM achieved state-of-the-art performance with a ROC-AUC of $99.7\%$ on the TSF dataset and demonstrated robust results on TF-66, a newly emerged, diverse, and privacy-preserving benchmark. These results highlight the generalizability and practicality of the proposed model, setting new standards for thermal fall detection and paving the way toward deployable, high-performance solutions.

Chee Mei Ling, Thangarajah Akilan, Aparna Ravinda Phalke

Agricultural image semantic segmentation is a pivotal component of modern agriculture, facilitating accurate visual data analysis to improve crop management, optimize resource utilization, and boost overall productivity. This study proposes an efficient image segmentation method for precision agriculture, focusing on accurately delineating farmland anomalies to support informed decision-making and proactive interventions. A novel Dual Atrous Separable Convolution (DAS Conv) module is integrated within the DeepLabV3-based segmentation framework. The DAS Conv module is meticulously designed to achieve an optimal balance between dilation rates and padding size, thereby enhancing model performance without compromising efficiency. The study also incorporates a strategic skip connection from an optimal stage in the encoder to the decoder to bolster the model's capacity to capture fine-grained spatial features. Despite its lower computational complexity, the proposed model outperforms its baseline and achieves performance comparable to highly complex transformer-based state-of-the-art (SOTA) models on the Agriculture Vision benchmark dataset. It achieves more than 66% improvement in efficiency when considering the trade-off between model complexity and performance, compared to the SOTA model. This study highlights an efficient and effective solution for improving semantic segmentation in remote sensing applications, offering a computationally lightweight model capable of high-quality performance in agricultural imagery.

Tushar Prasanna Swaminathan, Christopher Silver, Thangarajah Akilan

The proliferation of complex deep learning (DL) models has revolutionized various applications, including computer vision-based solutions, prompting their integration into real-time systems. However, the resource-intensive nature of these models poses challenges for deployment on low-computational power and low-memory devices, like embedded and edge devices. This work empirically investigates the optimization of such complex DL models to analyze their functionality on an embedded device, particularly on the NVIDIA Jetson Nano. It evaluates the effectiveness of the optimized models in terms of their inference speed for image classification and video action detection. The experimental results reveal that, on average, optimized models exhibit a 16.11% speed improvement over their non-optimized counterparts. This not only emphasizes the critical need to consider hardware constraints and environmental sustainability in model development and deployment but also underscores the pivotal role of model optimization in enabling the widespread deployment of AI-assisted technologies on resource-constrained computational systems. It also serves as proof that prioritizing hardware-specific model optimization leads to efficient and scalable solutions that substantially decrease energy consumption and carbon footprint.

Eduardo Reis, Thangarajah Akilan, Mohammed Khalid

Recent research in deep learning (DL) has investigated the use of the Fast Fourier Transform (FFT) to accelerate the computations involved in Convolutional Neural Networks (CNNs) by replacing spatial convolution with element-wise multiplications on the spectral domain. These approaches mainly rely on the FFT to reduce the number of operations, which can be further decreased by adopting the Real-Valued FFT. In this paper, we propose using the phasor form, a polar representation of complex numbers, as a more efficient alternative to the traditional approach. The experimental results, evaluated on the CIFAR-10, demonstrate that our method achieves superior speed improvements of up to a factor of 1.376 (average of 1.316) during training and up to 1.390 (average of 1.321) during inference when compared to the traditional rectangular form employed in modern CNN architectures. Similarly, when evaluated on the CIFAR-100, our method achieves superior speed improvements of up to a factor of 1.375 (average of 1.299) during training and up to 1.387 (average of 1.300) during inference. Most importantly, given the modular aspect of our approach, the proposed method can be applied to any existing convolution-based DL model without design changes.

Thangarajah Akilan

The coronavirus continues to disrupt our everyday lives as it spreads at an exponential rate. It needs to be detected quickly in order to quarantine positive patients so as to avoid further spread. This work proposes a new convolutional neural network (CNN) architecture called 'slow Encoding CNN. The proposed model's best performance wrt Sensitivity, Positive Predictive Value (PPV) found to be SP=0.67, PP=0.98, SN=0.96, and PN=0.52 on AI AGAINST COVID19 - Screening X-ray images for COVID-19 Infections competition's test data samples. SP and PP stand for the Sensitivity and PPV of the COVID-19 positive class, while PN and SN stand for the Sensitivity and PPV of the COVID-19 negative class.

Thangarajah Akilan, Edna Johnson, Japneet Sandhu et al.

Simultaneous localization and mapping (SLAM) is used to predict the dynamic motion path of a moving platform based on the location coordinates and the precise mapping of the physical environment. SLAM has great potential in augmented reality (AR), autonomous vehicles, viz. self-driving cars, drones, Autonomous navigation robots (ANR). This work introduces a hybrid learning model that explores beyond feature fusion and conducts a multimodal weight sewing strategy towards improving the performance of a baseline SLAM algorithm. It carries out weight enhancement of the front end feature extractor of the SLAM via mutation of different deep networks' top layers. At the same time, the trajectory predictions from independently trained models are amalgamated to refine the location detail. Thus, the integration of the aforesaid early and late fusion techniques under a hybrid learning framework minimizes the translation and rotation errors of the SLAM model. This study exploits some well-known deep learning (DL) architectures, including ResNet18, ResNet34, ResNet50, ResNet101, VGG16, VGG19, and AlexNet for experimental analysis. An extensive experimental analysis proves that hybrid learner (HL) achieves significantly better results than the unimodal approaches and multimodal approaches with early or late fusion strategies. Hence, it is found that the Apolloscape dataset taken in this work has never been used in the literature under SLAM with fusion techniques, which makes this work unique and insightful.

Apurva Pandya, Catherine Samuel, Nisargkumar Patel et al.

One of the most important tasks in medical image processing is the brain's whole tumor segmentation. It assists in quicker clinical assessment and early detection of brain tumors, which is crucial for lifesaving treatment procedures of patients. Because, brain tumors often can be malignant or benign, if they are detected at an early stage. A brain tumor is a collection or a mass of abnormal cells in the brain. The human skull encloses the brain very rigidly and any growth inside this restricted place can cause severe health issues. The detection of brain tumors requires careful and intricate analysis for surgical planning and treatment. Most physicians employ Magnetic Resonance Imaging (MRI) to diagnose such tumors. A manual diagnosis of the tumors using MRI is known to be time-consuming; approximately, it takes up to eighteen hours per sample. Thus, the automatic segmentation of tumors has become an optimal solution for this problem. Studies have shown that this technique provides better accuracy and it is faster than manual analysis resulting in patients receiving the treatment at the right time. Our research introduces an efficient strategy called Multi-channel MRI embedding to improve the result of deep learning-based tumor segmentation. The experimental analysis on the Brats-2019 dataset wrt the U-Net encoder-decoder (EnDec) model shows significant improvement. The embedding strategy surmounts the state-of-the-art approaches with an improvement of 2% without any timing overheads.

Yimin Yang, Q. M. Jonathan Wu, Xiexing Feng et al.

An iterative method of learning has become a paradigm for training deep convolutional neural networks (DCNN). However, utilizing a non-iterative learning strategy can accelerate the training process of the DCNN and surprisingly such approach has been rarely explored by the deep learning (DL) community. It motivates this paper to introduce a non-iterative learning strategy that eliminates the backpropagation (BP) at the top dense or fully connected (FC) layers of DCNN, resulting in, lower training time and higher performance. The proposed method exploits the Moore-Penrose Inverse to pull back the current residual error to each FC layer, generating well-generalized features. Then using the recomputed features, i.e., the new generalized features the weights of each FC layer is computed according to the Moore-Penrose Inverse. We evaluate the proposed approach on six widely accepted object recognition benchmark datasets: Scene-15, CIFAR-10, CIFAR-100, SUN-397, Places365, and ImageNet. The experimental results show that the proposed method obtains significant improvements over 30 state-of-the-art methods. Interestingly, it also indicates that any DCNN with the proposed method can provide better performance than the same network with its original training based on BP.

Thangarajah Akilan

Foreground (FG) pixel labelling plays a vital role in video surveillance. Recent engineering solutions have attempted to exploit the efficacy of deep learning (DL) models initially targeted for image classification to deal with FG pixel labelling. One major drawback of such strategy is the lacking delineation of visual objects when training samples are limited. To grapple with this issue, we introduce a multi-view receptive field fully convolutional neural network (MV-FCN) that harness recent seminal ideas, such as, fully convolutional structure, inception modules, and residual networking. Therefrom, we implement a system in an encoder-decoder fashion that subsumes a core and two complementary feature flow paths. The model exploits inception modules at early and late stages with three different sizes of receptive fields to capture invariance at various scales. The features learned in the encoding phase are fused with appropriate feature maps in the decoding phase through residual connections for achieving enhanced spatial representation. These multi-view receptive fields and residual feature connections are expected to yield highly generalized features for an accurate pixel-wise FG region identification. It is, then, trained with database specific exemplary segmentations to predict desired FG objects. The comparative experimental results on eleven benchmark datasets validate that the proposed model achieves very competitive performance with the prior- and state-of-the-art algorithms. We also report that how well a transfer learning approach can be useful to enhance the performance of our proposed MV-FCN.

Thangarajah Akilan, Q. M. Jonathan Wu, Wei Jiang

Following the rapidly growing digital image usage, automatic image categorization has become preeminent research area. It has broaden and adopted many algorithms from time to time, whereby multi-feature (generally, hand-engineered features) based image characterization comes handy to improve accuracy. Recently, in machine learning, pre-trained deep convolutional neural networks (DCNNs or ConvNets) have been that the features extracted through such DCNN can improve classification accuracy. Thence, in this paper, we further investigate a feature embedding strategy to exploit cues from multiple DCNNs. We derive a generalized feature space by embedding three different DCNN bottleneck features with weights respect to their Softmax cross-entropy loss. Test outcomes on six different object classification data-sets and an action classification data-set show that regardless of variation in image statistics and tasks the proposed multi-DCNN bottleneck feature fusion is well suited to image classification tasks and an effective complement of DCNN. The comparisons to existing fusion-based image classification approaches prove that the proposed method surmounts the state-of-the-art methods and produces competitive results with fully trained DCNNs as well.