Papers by Raafi Careem
Journal of Engineering in Industrial Research, 2025
This study presents a statistical evaluation of GRMobiNet, a novel lightweight deep neural networ... more This study presents a statistical evaluation of GRMobiNet, a novel lightweight deep neural network model designed for efficient image classification in resource-constrained environments (RCEs). Built upon the MobileNet architecture, GRMobiNet introduces targeted modifications to enhance predictive accuracy without increasing model size and computational complexity. To validate the performance gains, a series of controlled experiments were conducted using a custom dataset comprising ten image categories, with each model evaluated across ten repeated runs on identical test sets. A paired samples t-test was applied to compare the classification accuracy of GRMobiNet and MobileNetV2 under identical experimental conditions. Results indicate that GRMobiNet achieved a mean accuracy of 80%, significantly outperforming MobileNetV2’s 57%, with a mean improvement of 2.3 correctly predicted images per run. The observed p-value of 0.019 confirms statistical significance at the 95% confidence level. Moreover, GRMobiNet exhibited lower variance and a reduced standard error of the mean, indicating greater stability across trials. These findings confirm that GRMobiNet offers not only computational efficiency, but also statistically validated performance superiority, making it highly suitable for real-world deployment in domains such as mobile diagnostics, precision agriculture, and embedded surveillance systems. The statistical rigor of this validation underscores GRMobiNet’s robustness and reliability as an advancement over existing lightweight architectures.
Journal of Engineering in Industrial Research, 2024
This study aims to identify a baseline model for optimizing deep neural network (DNN) models for ... more This study aims to identify a baseline model for optimizing deep neural network (DNN) models for deployment in resource-constrained environments (RCE). Although DNNs are excellent in many applications, their deployment on devices like wearables and mobile phones presents significant challenges. The study investigates six popular DNN models, including MobileNet (V1 and V2), ResNet50, InceptionV3, DenseNet121, and EfficientNetB1. To assess each model's advantages, disadvantages, usability, and effectiveness in RCE scenarios, a comprehensive review and empirical analysis were conducted. The analysis focuses on optimizing these models to function effectively given the limited computational power and memory of RCE devices. Key factors such as model size, computational complexity, and inference speed are examined to uncover performance trade-offs between accuracy and resource efficiency. The findings suggest that MobileNetV1 should serve as baseline models for building efficiency-focused DNN models for image classification on RCE devices. This recommendation is based on MobileNetV1's balance between performance and efficiency, making it an ideal starting point for further optimization.
Indonesian Journal of Electrical Engineering and Computer Science, 2024
This paper aims to present a comprehensive review of advanced techniques and models with a specif... more This paper aims to present a comprehensive review of advanced techniques and models with a specific focus on deep neural network (DNN) for resource-constrained environments (RCE). The paper contributes by highlighting the RCE devices, analyzing challenges, reviewing a broad range of optimization techniques and DNN models, and offering a comparative assessment. The findings provide potential optimization techniques and recommend a baseline model for future development. It encompasses a broad range of DNN optimization techniques, including network pruning, weight quantization, knowledge distillation, depthwise separable convolution, residual connections, factorization, dense connections, and compound scaling. Moreover, the review analyzes the established optimization models which utilizes the above optimization techniques. A comprehensive analysis is conducted for each technique and model, considering its specific attributes, usability, strengths, and limitations in the context of effective deployment in RCEs. The review also presents a comparative assessment of advanced DNN models' deployment for image classification, employing key evaluation metrics such as accuracy and efficiency factors like memory and inference time. The article concludes with the finding that combining depthwise separable convolution, weight quantization, and pruning represents potential optimization techniques, while also recommending EfficientNetB1 as a baseline model for the future development of optimization models in RCE image classification.
International Journal of Computing and Digital Systems, 2024
This paper presents an empirical study on advanced Deep Neural Network (DNN) models, with a focus... more This paper presents an empirical study on advanced Deep Neural Network (DNN) models, with a focus on identifying potential baseline models for efficient deployment in resource-constrained environments (RCE). The systematic evaluation encompasses ten state-of-the-art pre-trained DNN models: ResNet50, InceptionResNetV2, InceptionV3, MobileNet, MobileNetV2, EfficientNetB0, EfficientNetB1, EfficientNetB2, DenseNet121, and Xception, within the context of an RCE setting. Evaluation criteria, such as parameters (indicating model complexity), storage space (reflecting storage requirements), CPU usage time (for real-time applications), and accuracy (reflecting prediction truth), are considered through systematic experimental procedures. The results highlight MobileNet's excellent trade-off between accuracy and resource requirements, especially in terms of CPU and storage consumption, in experimental scenarios where image predictions are performed on an RCE device. Utilizing the identified baseline model, a new model, GRM-MobileNet, was developed by implementing compound scaling and global average pooling techniques. GRM-MobileNet exhibits a substantial reduction of 23.81% in parameters compared to MobileNet, leading to a model size that is 23.88% smaller. Moreover, GRM-MobileNet demonstrates a significant improvement in accuracy, achieving a remarkable gain of 28.12% over MobileNet. Although the enhancement in inference time for GRM-MobileNet compared to MobileNet is modest at 1.66%, the overall improvements underscore the effectiveness of the employed strategies in enhancing the model's performance. A future study will examine other model optimization strategies, including factorization and pruning, which ultimately lead to faster inference without compromising accuracy, in an effort to improve the efficiency of the GRM-MobileNet model and its inference time.
Conference Presentations by Raafi Careem
International Conference on Computational and Mathematical Modelling , 2024
This paper presents a review and analysis of advanced deep neural network (DNN) models with a spe... more This paper presents a review and analysis of advanced deep neural network (DNN) models with a specific focus on optimizing them for the deployment in resource-constrained environments (RCE). Despite their outstanding performance in several applications, the implementation of DNNs on RCE devices, including mobile phones and wearables, poses challemges.
The review begins by assessing the suitability of RCE devices for image classification applications and delves into the distinctive challenges posed by the deployment of current DNN models on such devices. It evaluates established optimization models such as ResNet, MobileNet, InceptionNet, SqueezeNet, DenseNet, and EfficientNet.
Each model undergoes a thorough comprehensive and empirical analyses, considering its specific attributes, algorithms, usability, strengths, and limitations in the context of effective deployment in RCE scenarios. The review concludes to utilize the EfficientNetB1 as a baseline model to develop new optimization model for image classifications in RCE devices in future research.
Uploads
Papers by Raafi Careem
Conference Presentations by Raafi Careem
The review begins by assessing the suitability of RCE devices for image classification applications and delves into the distinctive challenges posed by the deployment of current DNN models on such devices. It evaluates established optimization models such as ResNet, MobileNet, InceptionNet, SqueezeNet, DenseNet, and EfficientNet.
Each model undergoes a thorough comprehensive and empirical analyses, considering its specific attributes, algorithms, usability, strengths, and limitations in the context of effective deployment in RCE scenarios. The review concludes to utilize the EfficientNetB1 as a baseline model to develop new optimization model for image classifications in RCE devices in future research.