The team is exploring opportunities to improve accuracy by revising the neural network architecture and fine-tuning hyper parameters. These adjustments aim to elevate the model's effectiveness in accurately recognizing emotions from voice data within the targeted use case.
Also to deal with an audio audio composed of both end speakers, the model has to be further trained to handle it.
Solutions recommended:
- Use tonal separation to differentiate the caller's voice and the employee's voice.
- Build a solution that records audio from the caller's end only.
The team is exploring opportunities to improve accuracy by revising the neural network architecture and fine-tuning hyper parameters. These adjustments aim to elevate the model's effectiveness in accurately recognizing emotions from voice data within the targeted use case.
Also to deal with an audio audio composed of both end speakers, the model has to be further trained to handle it.
Solutions recommended: