Institution profile

Khulna University

Academic institutionasia · bd
Official website
Research library3linked papers
Opportunities0open roles
Selected work

Representative Papers

Neural Network-based Study for Rice Leaf Disease Recognition and Classification: A Comparative Analysis Between Feature-based Model and Direct Imaging Model

Jul 03, 2025

Rice leaf diseases cause substantial yield losses, necessitating early and accurate identification. This study systematically compares two paradigms: Feature Analysis–based Detection Models (FADM) and Direct Image-Centered Detection Models (DICDM). FADM integrates multi-scale feature extraction, dimensionality reduction (e.g., PCA), feature selection (e.g., mRMR), and Extreme Learning Machine (ELM) classification, evaluated via 10-fold cross-validation; DICDM adopts an end-to-end image-input approach. Experimental results demonstrate that FADM achieves significantly higher classification accuracy and computational efficiency than DICDM across multiple rice disease classes. These findings validate the effectiveness of the “feature-driven” paradigm in resource-constrained agricultural settings. The work contributes a lightweight, interpretable, and empirically grounded framework for intelligent crop disease diagnosis, offering both methodological innovation and practical guidance for deploying AI in low-infrastructure farming environments.

0 citationsRead paper

Design of a 6-bit Threshold Inverter Quantization (TIQ) Flash Analog to Digital Converter (ADC)

Jan 08, 2025

Conventional flash ADCs suffer from process/temperature sensitivity and excessive area/power overhead due to resistor-divider networks and op-amp-based comparators. To address these limitations, this paper proposes a 6-bit fully digital flash ADC based on Threshold-Inversion Quantization (TIQ). The design replaces the conventional resistor ladder and op-amp comparators entirely with a TIQ comparator array built from two-stage CMOS inverters. A gain-enhancement circuit is integrated to improve resolution, while a 1-out-of-n encoder combined with a fat-tree architecture—novelly introduced for 6-bit flash ADCs—is employed to minimize encoding delay and area. Post-layout Tanner simulations in 0.25-μm CMOS demonstrate robust performance: at 2.5 V supply, power consumption is 6.25 mW with 1.07 μs latency for 10 kHz input; for 10 MHz input, it consumes 12.12 mW with 947.14 ns latency. The results validate the ADC’s high speed, low power, and strong process robustness.

0 citationsRead paper

Brain Controlled Wheelchair with Smart Feature

Jan 06, 2025

For severely paralyzed yet cognitively intact users, this study proposes a low-cost, high-reliability dual-modal brain–computer interface (BCI) powered wheelchair system. Methodologically, it introduces the first integration of EEG signals from the MindWave Mobile headset with eyelid-blink features to train a lightweight intent recognition model; concurrently, an Arduino-based platform fuses ultrasonic, inclinometer, and smoke sensor data to enable multi-source active safety responses, while an Android application—coupled with threshold-triggered SMS alerts—supports human–machine collaborative monitoring. Key contributions include: (1) blink-based navigation accuracy exceeding 92% and obstacle-avoidance response latency under 200 ms; (2) 100% automated detection and alerting for fall and smoke incidents; and (3) total system cost reduced to one-fifth that of conventional BCI wheelchairs, markedly enhancing clinical accessibility.

0 citationsRead paper
Recent publications

Latest Papers

Neural Network-based Study for Rice Leaf Disease Recognition and Classification: A Comparative Analysis Between Feature-based Model and Direct Imaging Model

Jul 03, 2025

Rice leaf diseases cause substantial yield losses, necessitating early and accurate identification. This study systematically compares two paradigms: Feature Analysis–based Detection Models (FADM) and Direct Image-Centered Detection Models (DICDM). FADM integrates multi-scale feature extraction, dimensionality reduction (e.g., PCA), feature selection (e.g., mRMR), and Extreme Learning Machine (ELM) classification, evaluated via 10-fold cross-validation; DICDM adopts an end-to-end image-input approach. Experimental results demonstrate that FADM achieves significantly higher classification accuracy and computational efficiency than DICDM across multiple rice disease classes. These findings validate the effectiveness of the “feature-driven” paradigm in resource-constrained agricultural settings. The work contributes a lightweight, interpretable, and empirically grounded framework for intelligent crop disease diagnosis, offering both methodological innovation and practical guidance for deploying AI in low-infrastructure farming environments.

0 citationsRead paper

Design of a 6-bit Threshold Inverter Quantization (TIQ) Flash Analog to Digital Converter (ADC)

Jan 08, 2025

Conventional flash ADCs suffer from process/temperature sensitivity and excessive area/power overhead due to resistor-divider networks and op-amp-based comparators. To address these limitations, this paper proposes a 6-bit fully digital flash ADC based on Threshold-Inversion Quantization (TIQ). The design replaces the conventional resistor ladder and op-amp comparators entirely with a TIQ comparator array built from two-stage CMOS inverters. A gain-enhancement circuit is integrated to improve resolution, while a 1-out-of-n encoder combined with a fat-tree architecture—novelly introduced for 6-bit flash ADCs—is employed to minimize encoding delay and area. Post-layout Tanner simulations in 0.25-μm CMOS demonstrate robust performance: at 2.5 V supply, power consumption is 6.25 mW with 1.07 μs latency for 10 kHz input; for 10 MHz input, it consumes 12.12 mW with 947.14 ns latency. The results validate the ADC’s high speed, low power, and strong process robustness.

0 citationsRead paper

Brain Controlled Wheelchair with Smart Feature

Jan 06, 2025

For severely paralyzed yet cognitively intact users, this study proposes a low-cost, high-reliability dual-modal brain–computer interface (BCI) powered wheelchair system. Methodologically, it introduces the first integration of EEG signals from the MindWave Mobile headset with eyelid-blink features to train a lightweight intent recognition model; concurrently, an Arduino-based platform fuses ultrasonic, inclinometer, and smoke sensor data to enable multi-source active safety responses, while an Android application—coupled with threshold-triggered SMS alerts—supports human–machine collaborative monitoring. Key contributions include: (1) blink-based navigation accuracy exceeding 92% and obstacle-avoidance response latency under 200 ms; (2) 100% automated detection and alerting for fall and smoke incidents; and (3) total system cost reduced to one-fifth that of conventional BCI wheelchairs, markedly enhancing clinical accessibility.

0 citationsRead paper