A Two-Pass Deep Learning System for Monitoring Visual Attributes of Food in Real-time during Fluidized Bed Drying
Bibliographic record
Abstract
Abstract Consumers rely on visual attributes when purchasing dried foods. If the product is unattractive, they walk away, leading to increase in global food waste. Attempts have been made to develop computer vision (CV) systems to monitor visual attributes of foods during the drying process. Unfortunately, figure-ground separation challenges such as overlapping, clustering, and variation in color and texture prevented the development of effective solutions for monitoring visual attributes of food during fluidized bed drying. To resolve this problem, we investigated the use of “Unet-Xception”, a novel real-time deep learning solution for monitoring the color, texture, and size of green peas, during fluidized bed drying. “Unet-Xception” combined modified U-Net and Xception architectures. U-Net segmented images of the peas, while Xception predicted visual attributes using the segmented output. Unet-Xception achieved a Mean Intersection-Over-Union of 0.9464 for segmentation quality, surpassing a classical CV solution. The AI-solution also detected additional objects of interest and outperformed the classical CV model in predicting visual attributes. It was found that a* and b* indices were the best predictors of color during drying. Homogeneity was the best parameter for monitoring texture. As expected, with improved segmentation and the detection of additional objects of interest, Unet-Xception produced far smoother trends than the classical model during deployment. This adaptable and novel solution is therefore able to monitor real-time changes in visual attributes of food, during fluidized bed drying. Incorporating this solution into current food dryers could lead to consistent product quality and significant reduction in global food waste.
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.
How this classification was reachedexpand
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 | 0.003 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".