Abhishek Mukherjee, PhD
Associate Professor
Agricultural and Ecological Research Unit
Biological Sciences Division
Indian Statistical Institute, Giridih Branch
Giridih, Jharkhand
I am an ecologist and entomologist interested in studying the ecology and management of invasive weed species in natural and agricultural systems. My broad areas of interest includes: ecological niche modelling of invasive weeds, plant – pest interactions, biorational management of important pest and diseases and applications of geostatistical technique to identify spatial patterns of species distribution.
Recent papers
A deep learning framework for species-level identification of rootknot nematodes using perineal pattern images
Sristi Das, Koyel Ghosh, Sandip Mondal, Suvasri Dutta, Matiyar Rahaman Khan, Utpal Garain & Abhishek Mukherjee
Scientific Report
Machine learning is emerging as an important tool in precision agriculture, including pest and disease diagnostics. This study developed CNN-based models using AlexNet and VGG16 to identify three economically important root-knot nematode species: Meloidogyne graminicola, M. incognita, and M. javanica. The model achieved about 95% accuracy, while manual identification by three annotators showed only moderate agreement (Kappa = 0.56), highlighting the limitations of expert-based identification. Integrated Gradients analysis showed that the model relied on taxonomically relevant perineal-pattern features, and cross-validation indicated stable performance with little overfitting. Overall, the study demonstrates that deep learning can provide rapid, accurate, and consistent nematode identification, especially useful where taxonomic expertise is limited.
Visualisation of model interpretability using integrated gradients for representative perineal pattern images of root-knot nematode species. Each row corresponds to one species: Meloidogyne graminicola, M. incognita, and M. javanica. For each species, the left image shows the original input image, the middle image shows the saliency map from standard gradients, and the right image shows the attribution map generated using the Integrated Gradients (IG) method. The highlighted regions (green contours) in the IG maps indicate the areas that most influenced the model’s classification. These regions broadly align with taxonomically important morphological features such as the dorsal arch and phasmid position.
