ScaledDenseNet: An Efficient Deep Learning Architecture for Skin Lesion Identification
Bibliographic record
Abstract
This research introduces ScaledDenseNet, a proficient deep learning architecture developed for precise identification of skin lesions.The model amalgamates DenseNet with the compound scaling method derived from EfficientNet, thereby achieving enhanced performance without detriment to the speed of inference.A grid search was conducted for the optimization of hyperparameters α, β, and Υ, instrumental in controlling the scaling of network dimensions.For model training and testing, the HAM10000 dataset was utilized, encompassing seven categories of diseases, namely carcinoma, basal cell carcinoma, benign keratosis-like lesions, dermatofibroma, melanoma, melanocytic nevi, and vascular lesions.ScaledDenseNet exhibited a top-3 accuracy of 94.59%.In response to the pronounced class imbalance within the dataset, image resizing was implemented to adjust input resolution based on phi for each architecture.A comparative analysis revealed that ScaledDenseNet surpassed DenseNet-121 and EfficientNet-B0, which achieved top-3 accuracies of 93.618% and 92.078%, respectively.The research methodology entailed a grid search for hyperparameter optimization and an explicit labeling scheme for disease categories, underpinning the study's validity and repeatability.Through its combination of DenseNet's extensive connectivity and the efficiency of the compound scaling method, ScaledDenseNet emerges as a promising tool for automated identification of skin lesions.Its performance underscores its potential applicability in the early detection and diagnosis of diverse skin conditions, marking a significant contribution to advancements in dermatological image analysis.
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 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 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".