Toward Fail-Safe Speaker Recognition: Trial-Based Calibration with a Reject Option
The output scores of most of the speaker recognition systems are not directly interpretable as stand-alone values. For this reason, a calibration step is usually performed on the scores to convert them into proper likelihood ratios, which have a clear probabilistic interpretation. The standard calib...
| Authors: | , , , , |
|---|---|
| Format: | article |
| Status: | Published version |
| Publication Date: | 2019 |
| Country: | Argentina |
| Institution: | Consejo Nacional de Investigaciones Científicas y Técnicas |
| Repository: | CONICET Digital (CONICET) |
| Language: | English |
| OAI Identifier: | oai:ri.conicet.gov.ar:11336/123318 |
| Online Access: | http://hdl.handle.net/11336/123318 |
| Access Level: | Open access |
| Keyword: | FORENSIC VOICE COMPARISON SPEAKER RECOGNITION TRIAL-BASED CALIBRATION https://purl.org/becyt/ford/1.2 https://purl.org/becyt/ford/1 |
| Summary: | The output scores of most of the speaker recognition systems are not directly interpretable as stand-alone values. For this reason, a calibration step is usually performed on the scores to convert them into proper likelihood ratios, which have a clear probabilistic interpretation. The standard calibration approach transforms the system scores using a linear function trained using data selected to closely match the evaluation conditions. This selection, though, is not feasible when the evaluation conditions are unknown. In previous work, we proposed a calibration approach for this scenario called trial-based calibration (TBC). TBC trains a separate calibration model for each test trial using data that is dynamically selected from a candidate training set to match the conditions of the trial. In this work, we extend the TBC method, proposing: 1) a new similarity metric for selecting training data that result in significant gains over the one proposed in the original work; 2) a new option that enables the system to reject a trial when not enough matched data are available for training the calibration model; and 3) the use of regularization to improve the robustness of the calibration models trained for each trial. We test the proposed algorithms on a development set composed of several conditions and on the Federal Bureau of Investigation multi-condition speaker recognition dataset, and we demonstrate that the proposed approach reduces calibration loss to values close to 0 for most of the conditions when matched calibration data are available for selection, and that it can reject most of the trials for which relevant calibration data are unavailable. |
|---|