Multimedia Data for the Visually Impaired
Authors: Niket Tandon, Shekhar Sharma, Tanima Makkad
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present preliminary results by conducting a user study with visually impaired people to measure the effectiveness of our system. Experiments Labeled data In order to construct labeled data, we conduct a user study with a group of 20 visually impaired people, over a dataset of 80 videos from various categories like news, games, speeches from Youtube. |
| Researcher Affiliation | Collaboration | Niket Tandon Max Planck Institute for Informatics Saarbr ucken, Germany ntandon@mpi-inf.mpg.de Shekhar Sharma PQRS Research pqrs-research.org shekhar.sharmaa@gmail.com Tanima Makkad PQRS Research pqrs-research.org tanima.vit@gmail.com |
| Pseudocode | Yes | Algorithm 1 Co-training algorithm Require: L: set of labeled, U: set of unlabeled points. Let h1 = textual feature based classifier, h2 = Visual feature based classifier, F1 = textual and F2 = visual features. 1: for K iterations do 2: Train classifiers h1 and h2 using F1 and F2 views of L resp., let h1 and h2 classify the instance in U. 3: Compute confidences of the classified instances in U with h1 s and h2s confidence measure. 4: If h1 classifies U with higher confidence, move labeled data to training set of h2 and vice-versa. 5: end for |
| Open Source Code | No | No explicit statement providing concrete access to source code for the methodology is found. The paper mentions 'Supplementary material at http://bit.ly/gyanjyoti' but does not specify it contains source code. |
| Open Datasets | No | The paper mentions a custom dataset: 'a dataset of 80 videos from various categories like news, games, speeches from Youtube' but does not provide concrete access information (link, DOI, repository, or formal citation for public availability). |
| Dataset Splits | No | The paper states '56 labeled videos were used as training set and rest as test set' but does not explicitly mention a validation set or its split information. |
| Hardware Specification | No | No specific hardware details (GPU/CPU models, processor types, or memory) used for running experiments were mentioned. |
| Software Dependencies | No | The paper mentions using 'SVM and Decision trees' for classification but does not provide specific version numbers for any software dependencies or libraries. |
| Experiment Setup | No | The paper describes the user study setup for data labeling and the overall dataset size, but it does not provide specific experimental setup details such as hyperparameters for SVM, Decision Trees, or the co-training algorithm (e.g., number of iterations or confidence thresholds). |