This paper introduces the DogAge dataset and an associated challenge to spark interest in the previously overlooked problem of automatic dog age estimation. While automatic age estimation using artificial neural networks like CNNs is a growing area for humans, it hasn’t been applied to dogs despite similarities in aging processes. The DogAge dataset, collected by animal and computer scientists, maps dog images to three age classes: young (0-2 years), adult (2-5 years), or senior (>6 years). It comprises expert data (1373 verified images) and Petfinder data (26190 unverified images from a pet adoption portal), which has been cleaned. The challenge proposes evaluating solutions using Mean Absolute Error (MAE), Average Recall (AR), and Categorization Accuracy (CA). The authors also provide ten baseline solutions combining CNN architectures (SqueezeNet and Inception v3) with various classifiers (kNN, SVM, Logistic Regression, Naive Bayes, Random Forest, feed-forward neural network) to serve as a starting point for research.
Non-Invasive Computer Vision-Based Fruit Fly Larvae Differentiation: Ceratitis capitata and Bactrocera zonata
This paper proposes a novel, non-invasive method using computer vision
