Hybrid Feature Fusion Using RNN and Pre-trained CNN for Classification of Alzheimer's Disease (Poster)
Bibliographic record
Abstract
The accurate classification of AD is very essential for both patient and social care, and it will be more significant once the treatment options are available to reverse the progress of the disease. The recent success of deep learning techniques has rapidly advanced the automatic classification of AD using neuroimaging biomarkers such as MRI. However, there exist two major challenges. First, training a deep convolutional neural network (CNN) from scratch relies on a large number of labeled training data to obtain high accuracy without overfitting. Second, due to high computational cost, most of the existing techniques employ 2D CNN that cannot leverage the complete spatial information; hence, it loses the inter-slice correlation. To address these limitations, we combine a recurrent neural network (RNN), specifically long short-term memory (LSTM) on top of the bottleneck layer of pre-trained DenseNet architecture, a deep CNN has already been trained on a large-scale dataset called ImageNet. In addition to the intra-slice features extracted from the deep CNN, the proposed technique exploits the inter-slice features through LSTM in order to discriminate the patients having AD and cognitively normal (CN) clinical status from the brain MRI data. Through experimental results, we show that our proposed model has better performance than state-of-the-art deep learning methods on the Open Access Series of Imaging Studies (OASIS) dataset using 5-fold cross validation.
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 | 0.001 |
| 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 | 0.001 | 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".