Optimizing Convolutional Neural Networks for Real-Time Object Detection in Autonomous Vehicles
Bibliographic record
Abstract
This study explores the optimization and reliability of convolutional neural network (CNN) models, namely YOLOv5, for real-time object detection in autonomous vehicles under varying environmental conditions. The study intends to evaluate the accuracy and speed of baseline models and optimized models based on the pruning, quantization, and knowledge distillation methods. With public datasets (KITTI, nuScenes, COCO) and local datasets from Islamabad and Karachi, the models were fine-tuned and trained to condition themselves to drive according to local driving conditions. Quantitatively, performance metrics such as mean Average Precision (mAP), frames per second (FPS), model size, and energy use were evaluated on high-end GPUs and embedded systems such as NVIDIA Jetson Xavier. Statistical tests such as paired t-tests and repeated measures ANOVA indicated that pruning decreases model size at the expense of slightly lower accuracy, quantization significantly accelerates inference while preserving good accuracy, and knowledge distillation achieves the optimal trade-off by retaining high accuracy and stability under harsh conditions such as low light, rain, and occlusion. These results emphasize the most essential trade-offs between efficiency and reliability in the deployment of deep models for autonomous vehicle perception systems. The research proposes using knowledge distillation via multi-condition learning and hardware-aware optimization to design reliable, real-time sufficient object detectors. This work establishes the practical deployment of efficient and reliable CNN models in autonomous vehicles for improved real-world performance and safety.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 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 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".