Ganesh Sistu

Ganesh Sistu

Auto-generated profile, based on publicly available information

Ganesh Sistu

Department of Electronic and Computer Engineering, University of Limerick, Limerick, Ireland

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 June 2025 in IEEE Open Journal of Vehicular Technology
Verified
Journal Article
Open access
MAMuhammad Asad
Muhammad Asad
IUIhsan Ullah
Ihsan Ullah
GSGanesh Sistu
Ganesh Sistu
MMMichael G. Madden
Michael G. Madden
In autonomous driving, understanding the surroundings is crucial for safety. Since most object detection systems are designed to identify known objects, they may miss unknown or novel objects, which can be dangerous. This study addresses Out-Of-Distribution (OOD) detection for vehicle-like unknown objects within the Bird's Eye View (BeV) space, a top-down representation of the environment that provides a comprehensive spatial layout crucial for scene understanding. Enhancing the model's adaptability to unfamiliar objects, we present two novel methods for detecting unknown objects in BeV space. Specifically, we introduce random patches and OOD objects in the environment to help the model identify both known objects, such as vehicles, and OOD objects. We also introduce a new dataset, NuScenesOOD, derived from the NuScenes dataset, which augments vehicles with patterns and shapes to challenge the model. Additionally, we address challenges such as patch size inconsistency and occlusion from moving frames in BeV space. Our method targets vehicle-shaped anomalies in the planar driving space, maintaining high accuracy for known and enhancing detection of unknown objects. This research contributes to making future autonomous vehicles safer by improving their ability to detect diverse vehicle like OOD objects in their environment. Code is available at: https://github.com/asad-python/OOD-in-BEV

13 February 2025 in IEEE Open Journal of Vehicular Technology
Verified
Journal Article
Open access
AMAnam Manzoor
Anam Manzoor
RMReenu Mohandas
Reenu Mohandas
ASAnthony G. Scanlan
Anthony G. Scanlan
EGEoin Martino Grua
Eoin Martino Grua
FCFiachra Collins
Fiachra Collins
CECiaran Eising
Ciaran Eising
The automotive industry has made significant strides in enhancing road safety and enabling automated driving features through advanced computer vision techniques. This is particularly true for short-range vehicle automation, where non-linear fisheye cameras are commonly used. However, these cameras are challenged by optical distortions, known as fisheye geometric distortions, which lead to object deformation within the image and significant pixel distortion, particularly at the image periphery. Based on the observation that fisheye and spherical images exhibit at least superficially similar geometric characteristics, we investigate the applicability of spherical models—including Spherical Convolutional Neural Networks (CNNs) and Spherical Vision Transformers (ViTs)—to fisheye images, even though fisheye images are not truly spherical. We perform our comparison using fisheye datasets—Woodscape, SynWoodscape, and SynCityscapes in autonomous driving scenarios, with a specific focus on the ability of spherical methods (Spherical CNNs and ViTs) to manage fisheye distortions and compared them against traditional non-spherical methods. Our findings indicate that spherical methods effectively address fisheye distortions without needing extra data augmentations. This results in better mean Intersection over Union (mIoU) scores, pixel accuracy, and better surround-view perception than other modern approaches for fisheye semantic segmentation. However, we also find that spherical methods have a greater tendency to overfit smaller datasets compared with non-spherical models. These advancements highlight how non-linear camera images can take advantage of spherical approximations through spherical models in autonomous driving.

5 June 2025 in IEEE Access
Verified
Journal Article
Open access
MAMuhammad Asad
Muhammad Asad
IUIhsan Ullah
Ihsan Ullah
MHMuhammad Adeel Hafeez
Muhammad Adeel Hafeez
GSGanesh Sistu
Ganesh Sistu
MMMichael G. Madden
Michael G. Madden
A novelty detection task involves identifying whether a data point is an outlier, given a training dataset that primarily captures the distribution of inliers. The novel class is usually absent, poorly sampled, or not well defined in the training data. A common technique for anomaly detection at present is to use an adversarial network generator to generate an anomaly score for inputs using the reconstruction loss. However, because this technique uses a competitive training process, it can be unreliable, with its performance being inconsistent during each adversarial training step. This inconsistency arises from changes in the network’s ability to detect anomalies. In this paper, we propose a revised framework for generative probabilistic novelty detection. We use a similar adversarial autoencoder-based framework but with a lightweight deep network, a novel training paradigm, and a probabilistic score to compute the reconstruction loss. Our methodology calculates the probability of whether a sample comes from the inlier distribution or not. The proposed approach can be applied to anomaly and outlier detection in images and videos. We present the results on multiple benchmark datasets, including the challenging UCSD Ped2 dataset for video anomaly detection. Our results illustrate that our proposed method learns the inlier classes and differentiates them from the outlier classes effectively, leading to better results than the baseline and state-of-the-art methods in several benchmark datasets.

5 November 2025 in IEEE Open Journal of Vehicular Technology
Verified
Journal Article
Open access
NTNikos Theodoridis
Nikos Theodoridis
TBTim Brophy
Tim Brophy
RMReenu Mohandas
Reenu Mohandas
GSGanesh Sistu
Ganesh Sistu
FCFiachra Collins
Fiachra Collins
CECiaran Eising
Ciaran Eising
Vision-Language Models (VLMs) are becoming increasingly powerful, demonstrating strong performance on tasks that require both visual and textual understanding. Their strong generalisation abilities make them a promising component for automated driving systems, which must handle unexpected corner cases. However, to be trusted in such safety-critical applications, a model must first possess a reliable perception system. Since critical objects and agents in traffic scenes are often at a distance, we require systems that are not “shortsighted,” i.e., systems with strong perception capabilities at both close (up to 20 meters) and long (30+ meters) range. With this in mind, we introduce Distance-Annotated Traffic Perception Question Answering (DTPQA), the first Visual Question Answering (VQA) benchmark focused solely on perception-based questions in traffic scenes, enriched with distance annotations. By excluding questions that require reasoning, we ensure that model performance reflects perception capabilities alone. Since automated driving hardware has limited processing power and cannot support large VLMs, our study centers on smaller VLMs. We evaluate several state-of-the-art (SOTA) small VLMs on DTPQA and show that, despite the simplicity of the questions, these models significantly underperform compared to humans (∼60% average accuracy for the best-performing small VLM versus ∼85% human performance). However, the human sample size was relatively small, which imposes statistical limitations. We also identify specific perception tasks, such as distinguishing left from right, that remain particularly challenging. We hope our findings will encourage further research into improving the perception capabilities of small VLMs in traffic scenarios, making them more suitable for automated driving applications.

30 April 2025 in IEEE Open Journal of Vehicular Technology
Verified
Journal Article
Open access
MHMena Nagiub Takwa HANNA
Mena Nagiub Takwa HANNA
TBThorsten Beuth
Thorsten Beuth
GSGanesh Sistu
Ganesh Sistu
HGHeinrich Gotzig
Heinrich Gotzig
CECiaran Eising
Ciaran Eising
This paper presents two approaches to enhance depth correction of the indirect time-of-flight (iToF) LiDAR sensors during a motion state, addressing the challenges of depth ambiguity and motion blur noise. iToF sensors are a key component in modern automotive applications, providing dense depth information for short-range vision applications for autonomous driving and Advanced Driver-Assistance Systems (ADAS). However, the periodic nature of iToF signals leads to depth ambiguity, making it challenging to measure distances accurately, especially in complex environments. Moreover, iToF sensors suffer from motion blur noise when the vehicle is in motion, compromising the accuracy of depth measurements. The proposed methods, which rely on depth correction indirectly through predicting depth bins using segmentation techniques, offer a promising alternative to direct depth regression. By focusing on segmentation-driven prediction, these new methods open up possibilities for more robust and precise depth correction in LiDAR sensor technology, potentially revolutionizing various applications that rely on accurate depth sensing. The results demonstrate the superiority of segmentation methods for depth frames based on 4 DCS samples, highlighting the potential impact and significance of this research in the field and the potential revolutionizing effect of these solutions on various applications.

23 February 2026 in IEEE Data Descriptions
Verified
Journal Article
Open access
SKSanjay Kumar
Sanjay Kumar
TBTim Brophy
Tim Brophy
RMReenu Mohandas
Reenu Mohandas
EGEoin Martino Grua
Eoin Martino Grua
GSGanesh Sistu
Ganesh Sistu
CECiaran Eising
Ciaran Eising
Robust perception in automated driving requires reliable performance under adverse conditions, where sensors may be affected by partial failures or environmental occlusions. Although existing automated driving datasets inherently contain sensor noise and environmental variability, very few enable controlled, parameterised, and reproducible degradations across multiple sensing modalities. This gap limits the ability to systematically evaluate how perception and fusion architectures perform under well-defined adverse conditions. To address this limitation, we introduce the OccNuScenes Dataset, a novel extension of the widely used nuScenes benchmark. For the camera modality, we release both the full and mini versions, each with four occlusion types: two adapted from public implementations and two newly designed. For radar and LiDAR, we provide parameterised occlusion scripts that implement three types of degradations each, enabling flexible and repeatable generation of occluded data. This resource supports consistent, reproducible evaluation of perception models under partial sensor failures and environmental interference. By releasing the first multi-sensor occlusion dataset with controlled and reproducible degradations, we aim to advance research on robust sensor fusion, resilience analysis, and safety-critical perception in automated driving.