Chickpea varietal classification using deep convolutional neural networks with transfer learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract The non‐availability of assessment tools among stakeholders often results in mixing of different chickpea varieties during its movement in the supply chain. Since each chickpea variety has unique physico‐chemical properties, it is important to prevent mixing with other varieties for maintaining the purity to obtain the intended quality for specific formulation. In this study, seven pre‐trained deep convolutional neural networks (AlexNet, GoogleNet, ResNet18, ResNet50, VGG16, VGG19, and MobileNetV2) with transfer learning were used for the classification of eight chickpea varieties (CDC‐Alma, CDC‐Leader, CDC‐Palmer, CDC‐Frontier, CDC‐Luna, CDC‐Orion, CDC‐Cory, and CDC‐Consul) using RGB images. For satisfying the input size requirement of the pre‐trained networks, the acquired images were cropped and resized using “Lanczos2” interpolation method for retaining maximum information from the original image. Furthermore, the hyperparameters of the pre‐trained networks (learning rate and batch size) were optimized to achieve high accuracy. The overall classification accuracy of the transfer learning models were 100, 100, 99, 92, 78, 72, and 50% for ResNet50, MobileNetV2, GoogleNet, ResNet18, VGG16, VGG19, and AlexNet, respectively. The study revealed that transfer learning is an effective way to derive the advantages of deep convolutional neural networks for varietal classification in chickpea. Practical Applications Classification of agricultural crops according to their varieties is critical during production and postharvest processing operations. Due to the growing importance of pulses as a major source of plant protein, the classification of chickpea on varietal basis assumes great significance to maintain the physico‐chemical characteristic uniqueness of the varieties. Hence, this study aimed at utilizing machine vision and deep transfer learning to successfully classify the different chickpea varieties. The developed model can be further integrated to a mobile environment or an embedded device for use in production catchments, processing facilities which would help the stakeholders in real time classification of chickpea varieties.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it