I recently came across an innovative use of AI image recognition, which details the use of a neural network to detect seal/cod worm parasites in codfish fillets. If you like sushi, it’s a queasy fact that fish can have live parasites, and these can infect humans. Most fish is deep-frozen to -20c to kill parasites and preserve freshness, so don’t be upset if you hear your sushi was previously frozen!
Here’s the abstract. Can you guess when this paper was written?
A neural-network-based system to recognize parasites/seal worms on cod fish images (McMaster University: http://hdl.handle.net/11375/24508)
A requirement for an intelligent system to automatically sort parasitic/sealworm infested cod fish fillet from good normal cod fish fillet has been recognized. 1 Hence, a neural-network-based system in combination with image processing techniques has been developed to recognize, detect and locate these parasites in cod fish fillets. This system combines the advantages of the massive parallelism of the neural network with image-processing procedures to recognize various attributes or characteristics of parasitic/sealworm patterns as derived from cod fish fillet images produced currently by backlighted photographs.
The generalized neural network backpropagation supervised learning algorithm is used for the investigation of the steepest descent and the conjugate gradient optimization methods.2 The cod fish fillet image data is manipulated so as to find the most efficient form, processed or unprocessed, before presentation to the neural network.
This paper reports on investigations and analysis carried out and then presents a proposed intelligent system to automatically separate parasitic/sealworm-infested cod fish fillets from normal cod fish fillets.