Object recognition devices, electronic devices and methods of recognizing objects

A method of recognizing an object in an object recognition device, which includes an artificial neural network engine, a nonvolatile memory device, and an error correction code (ECC) engine connected between the artificial neural network engine and the nonvolatile memory device, the method comprising:

• receiving learning data, weights and a feature vector from an outside;
• making an object recognition model learn by using the learning data and a selected portion of the weights;
• providing selected weight data, the selected weight data based on the selected portion of the weights;
• applying the feature vector and the selected weight data to the learned object recognition model to provide a recognition result of the object;
• performing an ECC encoding on the selected weight data to generate parity data;
• providing the selected weight data and the parity data to the nonvolatile memory device;
• providing the selected weight data to the artificial neural network engine by performing an ECC decoding on the selected weight data based on the parity data;
• and
• providing the selected weight data when a similarity between a result of applying the learning data and the selected weight data to the object recognition model and an expected value of the learning data is equal to or greater than a first reference value.