Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Mehedi Hasan has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Seeing Nothing, Saying Something: The Lack of Visual Grounding and Confabulation in Gemini Models for Histopathology

    by Md Mehedi Hasan, Mehmet Engin Tozal, Murat Seçkin Ayhan

    Large vision-language models (VLMs) have demonstrated remarkable performance on computational pathology benchmarks, yet their reliability under adversarial or vacuous inputs remains poorly understood. This paper examines the visual grounding behaviour of two Gemini models Gemini 3.0 Flash Preview ( gemini-flash ) and Gemini 3.1 Pro Preview ( gemini-pro ) on a well known histopathology classification task, and probes for confabulation using a adversarial blank-image set. On the real histopathology dataset both models achieve near-perfect accuracy (98.75% - 100%) across three temperatures (0.0, 0.5, 1.0) and three independent runs. On a controlled adversarial set of blank white images labelled as either benign or malignant, however, a stark divergence emerges. Gemini-flash consistently acknowledges the absence of visual content and assigns zero confidence, while Gemini-pro fabricates detailed, clinically plausible histological descriptions and reports high confidence (mean ≈ 0.95) across the same blank inputs, a behaviour we term confident confabulation . The confabulation rate of gemini-pro reaches 77.8% image-responses at temperature 0.0, dropping to 44.4% at temperature 0.5 and rising to 66.7% at temperature 1.0, while geminiflash records 0% at all temperatures. These findings raise important questions about the safety and trustworthiness of VLMs in clinical decision-support contexts, and underscore the need for comprehensive evaluation beyond standard accuracy metrics.

  • Top Cited Papers
    Open Access

    Potential smart grid vulnerabilities to cyber attacks: Current threats and existing mitigation strategies

    by Bishowjit Paul, Auvizit Sarker, Sarafat Hussain Abhi, Sajal Kumar Das, Firoj Ali, Manirul Islam, Robiul Islam, Sumaya Ishrat Moyeen, Faisal Rahman Badal, Hafiz Ahamed, et al.


  • Open Access

    Next generation legged robot locomotion: A review on control techniques

    by Swapnil Saha Kotha, Nipa Akter, Sarafat Hussain Abhi, Sajal Kumar Das, Robiul Islam, Firoj Ali, Hafiz Ahamed, Manirul Islam, Subrata Kumar Sarker, Faisal Rahman Badal, et al.


  • Top Cited Papers
    Open Access

    Digital twin: Data exploration, architecture, implementation and future

    by Shezad Dihan, Anwar Islam Akash, Zinat Tasneem, Prangon Das, Sajal Kumar Das, Robiul Islam, Manirul Islam, Faisal R. Badal, Firoj Ali, Hafiz Ahamed, et al.


  • Non-Invasive Portable Technologies for Monitoring Breast Cancer Related Lymphedema to Facilitate Telehealth: A Scoping Review

    by Arifur R. Anik, Kamrul Hasan, Manirul Islam, Mehedi Hasan, Firoj Ali, Sajal K. Das

    Breast cancer related lymphedema (BCRL) is a common, debilitating condition that can affect up to one in five breast cancer surviving patients (BCSP). BCRL can significantly reduce the quality of life (QOL) of patients and poses a significant challenge to healthcare providers. Early detection and continuous monitoring of lymphedema is crucial for the development of client-centered treatment plans for post-cancer surgery patients. Therefore, this comprehensive scoping review aimed to investigate the current technology methods used for the remote monitoring of BCRL and their potential to facilitate telehealth in the treatment of lymphedema. Initially, five electronic databases were systematically searched and analyzed following the PRISMA flow diagram. Studies were included, specifically if they provided data on the effectiveness of the intervention and were designed for the remote monitoring of BCRL. A total of 25 included studies reported 18 technological solutions to remotely monitor BCRL with significant methodological variation. Additionally, the technologies were categorized by method of detection and wearability. The findings of this comprehensive scoping review indicate that state-of-the-art commercial technologies were found to be more appropriate for clinical use than home monitoring, with portable 3D imaging tools being popular (SD 53.40) and accurate (correlation $>$ 0.9, p $< $ 0.05) for evaluating lymphedema in both clinic and home settings with expert practitioners and therapists. However, wearable technologies showed the most future potential for accessible and clinical long-term lymphedema management with positive telehealth outcomes. In conclusion, the absence of a viable telehealth device highlights the need for urgent research to develop a wearable device that can effectively track BCRL and facilitate remote monitoring, ultimately improving the quality of life for patients following post-cancer treatment.


  • Open Access

    TaLU: A Hybrid Activation Function Combining Tanh and Rectified Linear Unit to Enhance Neural Networks

    by Mehedi Hasan, Ali Hossain, Azmain Yakin Srizon, Abu Sayeed

    The application of the deep learning model in classification plays an important role in the accurate detection of the target objects. However, the accuracy is affected by the activation function in the hidden and output layer. In this paper, an activation function called TaLU, which is a combination of Tanh and Rectified Linear Units (ReLU), is used to improve the prediction. ReLU activation function is used by many deep learning researchers for its computational efficiency, ease of implementation, intuitive nature, etc. However, it suffers from a dying gradient problem. For instance, when the input is negative, its output is always zero because its gradient is zero. A number of researchers used different approaches to solve this issue. Some of the most notable are LeakyReLU, Softplus, Softsign, ELU, ThresholdedReLU, etc. This research developed TaLU, a modified activation function combining Tanh and ReLU, which mitigates the dying gradient problem of ReLU. The deep learning model with the proposed activation function was tested on MNIST and CIFAR-10, and it outperforms ReLU and some other studied activation functions in terms of accuracy(upto 6% in most cases, when used with Batch Normalization and a reasonable learning rate).

  • Open Access

    A State‐of‐the‐Art Computer Vision Adopting Non‐Euclidean Deep‐Learning Models

    by Sakib H. Chowdhury, Robius Sany, Hafiz Ahamed, Sajal K. Das, Faisal Rahman Badal, Prangon Das, Zinat Tasneem, Mehedi Hasan, Robiul Islam, Firoj Ali, et al.

    A distance metric known as non‐Euclidean distance deviates from the laws of Euclidean geometry, which is the geometry that governs most physical spaces. It is utilized when Euclidean distance is inappropriate, for as when dealing with curved surfaces or spaces with complex topologies. The ability to apply deep learning techniques to non‐Euclidean domains including graphs, manifolds, and point clouds is made possible by non‐Euclidean deep learning. The use of non‐Euclidean deep learning is rapidly expanding to study real‐world datasets that are intrinsically non‐Euclidean. Over the years, numerous novel techniques have been introduced, each with its benefits and drawbacks. This paper provides a categorized archive of non‐Euclidean approaches used in computer vision up to this point. It starts by outlining the context, pertinent information, and the development of the field’s history. Modern state‐of‐the‐art methods have been described briefly and categorized by application fields. It also highlights the model’s shortcomings in tables and graphs and shows different real‐world applicability. Overall, this work contributes to a collective information and performance comparison that will help enhance non‐Euclidean deep‐learning research and development in the future.


  • Open Access

    BengaliNet: A Low-Cost Novel Convolutional Neural Network for Bengali Handwritten Characters Recognition

    by Abu Sayeed, Jungpil Shin, Al Mehedi Hasan, Azmain Yakin Srizon, Mehedi Hasan

    As it is the seventh most-spoken language and fifth most-spoken native language in the world, the domain of Bengali handwritten character recognition has fascinated researchers for decades. Although other popular languages i.e., English, Chinese, Hindi, Spanish, etc. have received many contributions in the area of handwritten character recognition, Bengali has not received many noteworthy contributions in this domain because of the complex curvatures and similar writing fashions of Bengali characters. Previously, studies were conducted by using different approaches based on traditional learning, and deep learning. In this research, we proposed a low-cost novel convolutional neural network architecture for the recognition of Bengali characters with only 2.24 to 2.43 million parameters based on the number of output classes. We considered 8 different formations of CMATERdb datasets based on previous studies for the training phase. With experimental analysis, we showed that our proposed system outperformed previous works by a noteworthy margin for all 8 datasets. Moreover, we tested our trained models on other available Bengali characters datasets such as Ekush, BanglaLekha, and NumtaDB datasets. Our proposed architecture achieved 96–99% overall accuracies for these datasets as well. We believe our contributions will be beneficial for developing an automated high-performance recognition tool for Bengali handwritten characters.


  • Bangla Sign Digits: A Dataset For Real Time Hand Gesture Recognition

    by Dardina Tasmere, Boshir Ahmed, Mehedi Hasan

    With current widespread computer vision study, research with a particular culture and aesthetic contexts lead to enhanced performance with improved user experience. Despite a large number of deaf and dumb people, Bangla Sign Language has been less-focused on the research area. This paper proposes the building of identifying Bangla Sign Language using hand gestures to reduce the isolation of the hearing-impairment world. For the last little advancement in deep learning, Convolution Neural Network (CNN) has been used to classify Bangla Sign Language with higher accuracy. Our methodology overcomes all constraints from existing research of Bangla hand gesture recognition. In this paper, we presented a novel real-time method for Bangla sign digits (0-9) that focuses on three sections: image acquisition, preprocessing, and lastly, recognition of Bangla sign digits. For training, we have created a dataset of 1674 images from different people from various conditions. The proposed CNN model architecture classifies all Bangla sign digits with an overall accuracy of 97.63% on real-time video that ensures a potential gesture model that contributed significantly in the domain of Bangla sign number recognition compared to previous researches.