RESUMO
Audio classification techniques often depend on the availability of a large labeled training dataset for successful performance. However, in many application domains of audio classification (e.g., wildlife monitoring), obtaining labeled data is still a costly and laborious process. Motivated by this observation, a technique is proposed to efficiently learn a clean template from a few labeled, but likely corrupted (by noise and interferences), data samples. This learning can be done efficiently via tensorial dynamic time warping on the articulation index-based time-frequency representations of audio data. The learned template can then be used in audio classification following the standard template-based approach. Experimental results show that the proposed approach outperforms both (1) the recurrent neural network approach and (2) the state-of-the-art in the template-based approach on a wildlife detection application with few training samples.
Assuntos
Algoritmos , Animais Selvagens , Aprendizado de Máquina , Espectrografia do Som , Vocalização Animal , Animais , TempoRESUMO
Skeletal muscle injury resulting in tissue loss poses unique challenges for surgical repair. Despite the regenerative potential of skeletal muscle, if a significant amount of tissue is lost, skeletal myofibers will not grow to fill the injured area completely. Prior work in our lab has shown the potential to fill the void with an extracellular matrix (ECM) scaffold, resulting in restoration of morphology, but not functional recovery. To improve the functional outcome of the injured muscle, a muscle-derived ECM was implanted into a 1 x 1 cm(2), full-thickness defect in the lateral gastrocnemius (LGAS) of Lewis rats. Seven days later, bone-marrow-derived mesenchymal stem cells (MSCs) were injected directly into the implanted ECM. Partial functional recovery occurred over the course of 42 days when the LGAS was repaired with an MSC-seeded ECM producing 85.4 +/- 3.6% of the contralateral LGAS. This was significantly higher than earlier recovery time points (p < 0.05). The specific tension returned to 94 +/- 9% of the contralateral limb. The implanted MSC-seeded ECM had more blood vessels and regenerating skeletal myofibers than the ECM without cells (p < 0.05). The data suggest that the repair of a skeletal muscle defect injury by the implantation of a muscle-derived ECM seeded with MSCs can improve functional recovery after 42 days.