Annotation

Annotation is the process of labeling or marking the relevant features in a dataset. It involves identifying and outlining the objects or regions of interest that the model needs to learn and predict. Proper annotation is essential for training models in tasks such as object detection, instance segmentation, and semantic segmentation.

DeepLobe offers annotation capabilities to assist users in annotating their datasets. The platform provides an intuitive in-house annotation tool that enables users to mark and label objects, draw bounding boxes, create segmentation masks, and assign class labels to the annotated regions.

Annotations can be performed manually by users. DeepLobe facilitates a smooth annotation process, ensuring that the labeled data is accurately represented and ready for training the custom model. Users can invite annotators to take up the manual annotation tasks.

Model Metrics

Model metrics refer to the evaluation criteria used to assess the performance and effectiveness of a trained model. These metrics provide insights into how well the model is performing on specific tasks and help measure its accuracy, precision, recall, and other performance indicators.

When creating a custom model in DeepLobe, users can evaluate and monitor the model's performance using various metrics. These metrics are specific to the type of model and the task it is designed for. For instance:

1. Object Detection: Metrics such as mean Average Precision (mAP) and Intersection over Union (IoU) are commonly used to measure the accuracy and localization performance of object detection models.
2. Semantic Segmentation: Metrics like Pixel Accuracy, Mean Intersection over Union (mIoU), and Frequency Weighted Intersection over Union (FWIoU) are used to evaluate the pixel-wise segmentation accuracy.
3. Classification: Metrics such as accuracy, precision, recall, and F1 score can be used to evaluate the performance of classification models.

DeepLobe provides a model metrics dashboard that allows users to track and analyze these metrics during the training process. This helps users make informed decisions about model improvement, fine-tuning, and selecting the best-performing model for their specific use case.

By understanding and monitoring model metrics, users can iterate and enhance their models to achieve optimal performance and accuracy.

Note: The availability and applicability of specific metrics may vary depending on the type of model and the task at hand.