- 8th October: List of confirmed speakers is here. More to follow!
- 8th September: Our challenges are launched!
- 28th June: Challenges of FGIA will be held in conjunction with ACCV 2022. More to follow!
Fine-grained image analysis (FGIA) is a longstanding and fundamental problem in computer vision and pattern recognition, and underpins a diverse set of real-world applications. The task of FGIA targets analyzing visual objects from subordinate categories, e.g., species of birds or models of cars. The small inter-class and large intra-class variation inherent to fine-grained image analysis makes it a challenging problem. In our workshop, we would like to host series of academic challenges about fine-grained image analysis, including webly-supervised fine-grained recognition and large-scale fine-grained hashing search.
For the webly-supervised fine-grained recognition problem, participants have to deal with the challenging subordinate fine-grained category classification task (totally 5,089 sub-categories) by leveraging the free but noisy web data (a.k.a. webly fine-grained images). This results in a more robust evaluation of the current state of the art, counteracts overfitting to traditional fine-grained benchmarks, and allows insights into prevalent dataset bias, data noises and as well as long-tailed distribution of the webly fine-grained nature. The overall goal of the challenge is to evaluate the robustness and generalization capabilities of state-of-the-art algorithms.
The large-scale fine-grained hashing search problem refers to the task of learning binary hashing codes for large-scale fine-grained image retrieval. It is desirable to generate compact hash bits for fine-grained images sharing both large intra-class variances and small inter-class variances. Participants have to provide low dimensional hash codes for all the images in the test set (totally 1,000 sub-categories) by leveraging the provided training and valid data obtained from iNaturalist. The overall goal of this challenge is to evaluate the retrieval performance of those state-of-the-art algorithms associated with retrieval and hashing, which helps greatly reduce the storage costs and increase the query speeds.
Beyond these challenges, an invited talk session will be also considered. In such a session, the winner solutions will be invited to share the solutions of the challenges. Also, some experts in the related fields are invited to share insights and thinking about the main topic of our workshop.