Hey, I’m Rashid.

I’m a Machine Learning (Computer Vision) Researcher at the University of Turin in Italy, where I have been pursuing my PhD since 2022. My research is centered on advancing machine learning techniques for Computer vision related tasks, with a recent focus on eXplainable AI (XAI) and visual Anomaly Detection in industrial applications.

Objective

Skilled R&D expert in Computer Vision and eXplainable AI (XAI) specializing in Precise object localization using eXplainable AI and Visual Anomaly Detection. My previous work spans object detection, video surveillance, and computer vision aided medical imaging disease recognition.

My primary research interests include:

• Computer Vision
• Deep Learning
• Anomaly Detection
• Explainable Artificial Intelligence (XAI)
• Explainable Copmuter Vision (XCV)

News

🌳 ShapBPT: Image Feature Attributions using Data-Aware Binary Partition Trees @ AAAI-26

Our paper was accepted in the main technical track of the 40th Annual AAAI Conference on Artificial Intelligence (AAAI-26) and presented as a poster contribution.

Existing image-based eXplainable AI (XAI) methods often rely on fixed or heuristic feature partitions (e.g., superpixels or uniform grids), which may fail to align with the underlying data distribution and object structure. This limitation can lead to explanations that are either fragmented or semantically misaligned with the true decision-making process of the model.

ShapBPT introduces a novel XAI framework that integrates Data-Aware Binary Partition Trees (BPT) into Shapley-value-based feature attribution. By constructing hierarchical, data-driven partitions, the proposed method generates explanations that are both faithful to the model behavior and better aligned with meaningful image regions. Extensive experiments across multiple vision models and datasets demonstrate that ShapBPT consistently improves localization accuracy, stability, and human interpretability compared to state-of-the-art explanation techniques.

Method Availability

The method is available under: 👉 https://github.com/amparore/shap_bpt

Links: · Method · Tests & Experiments · Poster · Technical Appendix · PyPI Package - pip install shap-bpt

🤖 Can I Trust My Anomaly Detection System? A Case Study Based on Explainable AI @ XAI-World-24

• Our paper was accepted in main technical track and was presented in special session Explainable AI for improved human-computer interaction.

• Generative models based on variational autoencoders are a popular technique for detecting anomalies in images in a semi-supervised context. A common approach employs the anomaly score to detect the presence of anomalies, and it is known to reach high level of accuracy on benchmark datasets. However, since anomaly scores are computed from reconstruction disparities, they often obscure the detection of various spurious features, raising concerns regarding their actual efficacy. This case study explores the robustness of an anomaly detection system based on variational autoencoder generative models through the use of eXplainable AI methods. The goal is to get a different perspective on the real performances of anomaly detectors that use reconstruction differences. In our case study we discovered that, in many cases, samples are detected as anomalous for the wrong or misleading factors.

• Method Availability

  • The method is available under: https://github.com/rashidrao-pk/anomaly_detection_trust_case_study
  • Links: GitHub Paper PDF on Arxiv Proceedings of XAI, Download Slides

🤖 Using Stratified Sampling to Improve LIME Image Explanations @ AAAI-24

• Our paper was accepted in main technical track and then as a poster presentation, only 100 random papers were selected out of 2200 accepted papers and all the papers also required to be presented as poster presentation as well including those 100 papers.

• We investigate the use of a stratified sampling approach for LIME Image, a popular model-agnostic explainable AI method for computer vision tasks, in order to reduce the artifacts generated by typical Monte Carlo sampling. Such artifacts are due to the undersampling of the dependent variable in the synthetic neighborhood around the image being explained, which may result in inadequate explanations due to the impossibility of fitting a linear regressor on the sampled data. We then highlight a connection with the Shapley theory, where similar arguments about undersampling and sample relevance were suggested in the past. We derive all the formulas and adjustment factors required for an unbiased stratified sampling estimator. Experiments show the efficacy of the proposed approach.

• Method Availability

  • The method is available under: https://github.com/rashidrao-pk/lime_stratified
  • Examples and full results given at Lime_stratified_examples
  • Links: GitHub, Paper PDF on Arxiv Proceedings of AAAI-24

Publication List

First Author Publications

1. ShapBPT: Image Feature Attributions using Data-Aware Binary Partition Trees
   Muhammad Rashid, Elvio Amparore, Enrico Ferrari, Damiano Verda
   AAAI-26

2. Can I Trust My Anomaly Detection System? A Case Study Based on Explainable AI
   Muhammad Rashid, Elvio Amparore, Enrico Ferrari, Damiano Verda
   eXplainable AI world Conference-24

3. Using Stratified Sampling to Improve LIME Image Explanations
   Muhammad Rashid, Elvio Amparore, Enrico Ferrari, Damiano Verda
   AAAI-24

4. A sustainable deep learning framework for object recognition using multi-layers deep features fusion and selection
   Muhammad Rashid, Muhammad Attique Khan, Majed Alhaisoni, Shui-Hua Wang, Syed Rameez Naqvi, Amjad Rehman, Tanzila Saba
   Sustainability | Q1 Journal 3.25 IF

5. Object Detection and Classification: A Joint Selection and Fusion Strategy of Deep Convolutional Neural Network and SIFT Point Features
   Muhammad Rashid, Muhammad Attique Khan, Muhammad Sharif, Mudassar Raza, Muhammad Masood, Farhat Afza
   Multimedia Tools and Applications | Q1 Journal 2.577 IF

Further Publications

1. A Novel Light U-Net Model for Left Ventricle Segmentation Using MRI
   Mehreen Irshad, Mussarat Yasmin, Muhammad Imran Sharif, Muhammad Rashid, …
   Mathematics | Q Journal 2.3 IF

2. Deep CNN and geometric features-based gastrointestinal tract diseases detection and classification from wireless capsule endoscopy images
   Muhammad Sharif, Muhammad Attique Khan, Muhammad Rashid, Mussarat Yasmin, Farhat Afza, Urcun John Tanik
   Journal of Experimental & Theoretical Artificial Intelligence | Journal IF 2.296

3. Classification of gastrointestinal diseases of stomach from WCE using improved saliency-based method and discriminant features selection
   Muhammad Attique Khan, Muhammad Rashid , Muhammad Sharif, Kashif Javed
   Multimedia Tools and Applications - IF 2.577
4. An integrated framework of skin lesion detection and recognition through saliency method and optimal deep neural network features selection
   M Attique Khan, Tallha Akram, Muhammad Sharif, Kashif Javed, Muhammad Rashid, Syed Ahmad Chan Bukhari
   Neural Computing and Applications - IF 5.102
5. Region-based active contour JSEG fusion technique for skin lesion segmentation from dermoscopic images
   Rabia Javed, Mohd Shafry Mohd Rahim, Tanzila Saba, Muhammad Rashid
   Biomedical Research | Journal IF 0.219
6. An Optimized Approach for Breast Cancer Classification for Histopathological Images Based on Hybrid Feature Set
   Inzamam Mashood Nasir, Muhammad Rashid, Jamal Hussain Shah, Muhammad Sharif, Muhammad Yahiya Haider Awan, Monagi H Alkinani
   Current medical imaging | Journal IF 1.315

GitHub Stats