Automatic interpretation of salmon scales using deep learning

For several fish species, age and other important biological information is manually inferred from visual scrutinization of scales, and reliable automatic methods are not widely available. Here, we apply Convolutional Neural Networks (CNN) with transfer learning on a novel dataset of 9056 images of Atlantic salmon scales for four different prediction tasks. We predicted fish origin (wild/farmed), spawning history (previous spawner/nonspawner), river age, and sea age. We obtained high prediction accuracy for fish origin (96.70%), spawning history (96.40%), and sea age (86.99%), but lower accuracy for river age (63.20%). Against six human expert readers with an additional dataset of 150 scales, the CNN showed the second-highest percentage agreement for sea age (94.00%, range 87.25 +/- 97.30%), but the lowest agreement for river age (66.00%, range 66.00- 84.68%). Estimates of river age by expert readers exhibited higher variance and lower levels of agreement compared to sea age and may indicate why this task is also more difficult for the CNN. Automatic interpretation of scales may provide a cost- and time-efficient method of predicting fish age and life-history traits.

Automated summary

The study applied Convolutional Neural Networks with transfer learning on a dataset of Atlantic salmon scales, achieving high accuracy in predicting fish origin, spawning history, and sea age, but lower accuracy for river age. Comparison with human expert readers showed higher agreement in sea age prediction and lower agreement in river age, indicating the difficulty of the task.