Browsing by Author "Jahan, Ismat"

Now showing 1 - 4 of 4

A coarse-to-fine hierarchical framework for bone marrow cell recognition: integrating morphological features with class-specific augmentation and multi-perspective explainable AI
(BRAC University, 2025-10) Mamun, Abdullah AL; Haider, Zarin Tasnim; Montaha, Sidratul; Jahan, Ismat; Mukta, Jannatun Noor; Nasim, Hamim Ibne
The classification of bone marrow (BM) cell is essential for the diagnosis of many haematological disorders. Automated cytological analysis still suffers from extreme class imbalance, very high morphological similarity between cell types, and poor interpretability of most artificial intelligence (AI) models, despite advances in medical imaging and deep learning. We present an interpretable, state-of-the-art framework for BM cell classification based on a large-scale dataset with 171,374 single-cell images annotated by experts with 21 classes. Given the high severity of the class imbalance (originally 3678:1 at times), we created a new subset of 95,865 images from the overall dataset through segmentation and feature extraction. Through this process, overrepresented classes were down-sampled, while specific types of augmentation were applied to underrepresented classes to restore balance, resulting in a ratio of 1435:1. Our recursive segmentation approach, based on CMYK (Cyan, Magenta, Yellow, and Black) and HLS (Hue, Saturation, and Lightness) colour spaces, reliably identifies nucleus, cytoplasm, and whole-cell regions. From these areas, we processed 144 biologically motivated shape, colour, texture, and fractal attributes. We build a hierarchical two-stage classification model named HierEff-S2, where an EfficientNet-B4 backbone assigns each cell to one of six morphological groups. Then, group-specific EfficientNet-B3 models perform fine-grained classification within each group. With 21 classes, this architecture obtains 86.1% accuracy and outperforms other models, including VisionMamba, Ensemble Model, and MobileNet. To promote clinical interpretability, we combine two explainable AI methods to visually highlight cell regions that lead to the model predictions: Grad-CAM and LIME. Using XAI, we report 95.2% correctness at the image level, thus providing biologically meaningful attention to the model.
Neonate and Infant Mortality Rate; A Review
(6/10/2016) Jahan, Ismat
Childhood mortality, particularly in the first 5 years of life, is a major global concern and the target of Millennium Development Goal 4.Globally 7.6 million children died in 2010 before reaching their fifth birthday and 40% of these deaths occur in the neonatal period. Timely measurements of levels and trends in under-5 mortality are important to assess progress towards the Millennium Development Goal 4 (MDG 4) target of reduction of child mortality by two thirds from 1990 to 2013, and to identify models of success. To implement evidence-based interventions for the reduction of neonatal mortality, it is important to investigate factors associated with neonatal mortality. The aim of the current study was to identify determinants of neonatal mortality.To generated updated estimates of child mortality in early neonatal (age 0–6 days), late neonatal (7–28 days), postneonatal (29–364 days), childhood (1–4 years), and under-5 (0–4 years) age groups for some countries from 1970 to 2013. To quantify the contribution of these different factors and birth numbers to the change in numbers of deaths in under-5 age groups from 1990 to 2013.The study showed that the infantmortality rate (IMR), neonate mortality rate (NMR)& under five mortality rate (U5MR) for singleton live born infants between 1990-2013.Neonatal mortality has declined in all world regions. Progress has been slowest in the regions with high neonate mortality rates(NMRs). Global health programs need to address neonatal deaths moreeffectively if Millennium Development Goal 4 (two-thirds reduction in child mortality) is to be achieved.
TAXONOMIC CONFIRMATION AND SPECIES COMPOSITION OF MUD CRAB Scylla spp. AVAILABLE IN BANGLADESH
(A thesis submitted in the partial fulfillment of the requirements for the degree of Master of Science in Marine Bioresource Science Department of Marine Bioresource Science Faculty of Fisheries Chittagong Veterinary and Animal Sciences University Chittagong-4225, Bangladesh., 2018-06) Jahan, Ismat
Bangladesh is riverine country located in South Asia along with a coastline about 710 km including 618,780 ha mangrove with tidal flats which are greatly suitable for distribution and good composition for mud crab population. In past many researchers worked on the mud crab for aquaculture purpose in the real field without knowing their proper species recognition including with their composition in the coastal regions of Bangladesh. Recently some researchers dealt with taxonomic work on crab for species clarification based on molecular techniques which required highly modern equipped laboratory that is known to be difficult for general people including illiterate fishermen. This research was conducted for measuring species composition of the mud crab (Scylla spp.), in the particular area of Bangladesh. The present study was conducted at Bagerhat, Cox’s Bazar and Chittagong districts of Bangladesh for taxonomic confirmation and species composition of mud crab (Scylla). About 70-100 mud crab specimens were collected from each of those regions. From primary identification, 46%, 59% and 75% crabs were considered as S. olivacea, while 54%, 41%, and 25% crabs were S. serrata at Bagerhat, Cox’s Bazar and Chittagong, respectively. For more clarification, all crabs were observed external morphology including with the description of first and second male gonopods and measured 24 morphological characters for morphologically distinct samples. From the measured characters were calculated into 27 morphometric ratio to recognise each single species. The present study confirmed the existence of two species of mud crabs namely S. olivacea and S. serrata where 59% was S. olivacea and 41% was S. serrata from the collected sample in Bangladesh. The study revealed that S. olivacea is dominant mud crab species among four Scylla spp. after that S. serrata also well distributed in the coastal regions of Bangladesh.