Toward Energy-Efficient Deep Neural Networks for Forest Fire Detection in an Image
Bibliographic record
Abstract
Forest fires cause huge losses and are a serious problem facing many countries worldwide, including the USA, Canada, Brazil, Siberia, and Indonesia, to name a few. Automatic identification of forest fires in an image is thus an important field to research in order to minimize disasters while also helping in mitigation planning and designing rescue tactics. Artificial Intelligence technologies, especially deep neural networks, have emerged recently with promises to detect fires with better accuracy from an image. However, the massive energy consumption of deep neural networks thwarts their widespread adoption, especially when it comes to onsite detection of fire utilizing low-power devices such as those embedded in a drone or an artificial satellite. In this paper, we develop multiple deep neural network models such as a Convolutional Neural Network (CNN), a Deep Belief Network (DBN), an Auto Encoder (AEnc), and a U-net model to detect forest fires and systematically analyze their accuracy and energy consumption using IEEE FLAME Dateset which is openly available at IEEE data portal. After developing the models, we systematically pruned the models, retrained them, and analyzed their accuracy and energy consumption upon deployment. Our analysis shows that the CNN has the highest accuracy (almost 99%) on the validation data set, whereas the DBN model consumes the least amount of energy after deploying on both CPU and GPU. The trained models are deployed on a website for use. The source code can be found on GitHub (https://github.com/akdasUAF/ForestFireDetection).
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".