Attention-Based Deep Neural Network Combined Local and Global Features for Indoor Scene Recognition
Bibliographic record
Abstract
An original attention-based indoor scene recognition model combining local and global features is proposed. Multi-strategy data augmentation using several different functions and intensities can improve the classification performance. Then, local features are extracted using a convolutional layer and a single self-attention, thus solving the problem of large intra-class variance. The multi-attention mechanism is used to fuse the local feature information extracted from different foci to obtain a more complete global feature representation. The multi-head attention mechanism allows the network to extract features in parallel in different directions of attention, which helps the network to better capture global information, improves the network's ability to understand and represent the input data, and solves the problem of high inter-class similarity. Finally, the extracted features are fed into the classifier to complete the classification of indoor scene images. Experiments are conducted on four data sets (IndoorCVPR09, SUN397, 15-Scenes and self-built small sample scientific indoor scene dataset), yield excellent results. The results show that the developed algorithm effectively solves the two problems of high intra-class diversity and high inter-class similarity. As a result, the model has achieved competitive results. Preliminary application experiments are developed in our HRI system, indicating that the proposed indoor scene recognition model can be applied to the complete environmental perception in HRI.
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) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 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".