This paper proposes a thermal-infrared (TIR) remote target detection system for maritime rescue using deep learning and data augmentation. We established a self-collected TIR dataset consisting of multiple scenes imitating human rescue situations using a TIR camera (FLIR). Additionally, to address dataset scarcity and improve model robustness, a synthetic dataset from a 3D game (ARMA3) to augment the data is further collected. However, a significant domain gap exists between synthetic TIR and real TIR images. Hence, a proper domain adaptation algorithm is essential to overcome the gap. Therefore, we suggest a domain adaptation algorithm in a target-background separated manner from 3D game-to-real, based on a generative model, to address this issue. Furthermore, a segmentation network with fixed-weight kernels at the head is proposed to improve the signal-to-noise ratio (SNR) and provide weak attention, as remote TIR targets inherently suffer from unclear boundaries. Experiment results reveal that the network trained on augmented data consisting of translated synthetic and real TIR data outperforms that trained on only real TIR data by a large margin. Furthermore, the proposed segmentation model surpasses the performance of state-of-the-art segmentation methods.
Sound event detection (SED) is one of tasks to automate function by human auditory system which listens and understands auditory scenes. Therefore, we were inspired to make SED recognize sound events in the way human auditory system does. Spectro-temporal receptive field (STRF), an approach to describe the relationship between perceived sound at ear and transformed neural response in the auditory cortex, is closely related to recognition of sound. In this work, we utilized STRF as a kernel of the first convolutional layer in SED model to extract neural response from input sound to make SED model similar to human auditory system. In addition, we constructed two-branched SED model named as Two Branch STRFNet (TB-STRFNet) composed of STRF branch and baseline branch. While STRF branch extracts sound event information from auditory neural response, baseline branch extracts sound event information directly from the mel spectrogram just as conventional SED models do. TB-STRFNet outperformed the DCASE baseline by 4.3% in terms of threshold-independent macro F1 score, achieving 4th rank in DCASE Challenge 2023 Task 4b. We further improved TB-STRFNet by applying frequency dynamic convolution (FDYConv) which also leveraged domain knowledge on acoustics. As a result, two branch model applied with FDYConv on both branches outperformed the DCASE baseline by 6.2% in terms of the same metric.
We explore on various attention methods on frequency and channel dimensions for sound event detection (SED) in order to enhance performance with minimal increase in computational cost while leveraging domain knowledge to address the frequency dimension of audio data. We have introduced frequency dynamic convolution in a previous work to release the translational equivariance issue associated with 2D convolution on the frequency dimension of 2D audio data. Although this approach demonstrated state-of-the-art SED performance, it resulted in 2.5 times heavier model in terms of the number of parameters. To achieve comparable SED performance with computationally efficient methods to enhance practicality, we explore on lighter alternative attention methods. In addition, we focus of attention methods on frequency and channel dimensions as those are shown to be critical in SED. Joint application of SE modules on both frequency and channel dimension shows comparable performance to frequency dynamic convolution with only 2.7% increase in the model size compared to the baseline model. In addition, we performed class-wise comparison of various attention methods to further discuss their characteristics.
The goal of DCASE 2023 Challenge Task 7 is to generate various sound clips for Foley sound synthesis (FSS) by "category-to-sound" approach. "Category" is expressed by a single index while corresponding "sound" covers diverse and different sound examples. To generate diverse sounds for a given category, we adopt VITS, a text-to-speech (TTS) model with variational inference. In addition, we apply various techniques from speech synthesis including PhaseAug and Avocodo. Different from TTS models which generate short pronunciation from phonemes and speaker identity, the category-to-sound problem requires generating diverse sounds just from a category index. To compensate for the difference while maintaining consistency within each audio clip, we heavily modified the prior encoder to enhance consistency with posterior latent variables. This introduced additional Gaussian on the prior encoder which promotes variance within the category. With these modifications, we propose VIFS, variational inference for end-to-end Foley sound synthesis, which generates diverse high-quality sounds.
In this paper, an automatic calibration algorithm is proposed to reduce the depth error caused by internal stray light in amplitude-modulated continuous wave (AMCW) coaxial scanning light detection and ranging (LiDAR). Assuming that the internal stray light inside the proposed AMCW scanning LiDAR is static, the amplitude and phase delay of internal stray light are estimated using the Gaussian mixture model (GMM) and particle swarm optimization (PSO). Specifically, the raw amplitude (amplitude of raw cross-correlation) map of calibration checkboard at specific distance is segmented by GMM with two clusters (the bright and dark regions). The loss function is then defined as L1-norm of difference between mean depth of the amplitude-segmented clusters. This loss function is minimized by PSO to estimate the two optimal target parameters: the amplitude and phase delay of internal stray light. To avoid overfitting at a specific distance, the calibration check board is measured at multiple distances and the average of L1 loss functions from all measured distances is chosen as the actual loss. According to the validation of the proposed algorithm, the original loss is reduced from tens of centimeters to 3.2 mm when the distance of the calibration checkboard is between 1m and 4 m. This accurate depth error correction performance is also maintained in the depth and raw amplitude images of geometrically complex objects. The proposed internal stray light calibration algorithm in this paper can be used for any type of AMCW coaxial scanning LiDAR regardless of its optical characteristics.
Performance of sound event localization and detection (SELD) in real scenes is limited by small size of SELD dataset, due to difficulty in obtaining sufficient amount of realistic multi-channel audio data recordings with accurate label. We used two main strategies to solve problems arising from the small real SELD dataset. First, we applied various data augmentation methods on all data dimensions: channel, frequency and time. We also propose original data augmentation method named Moderate Mixup in order to simulate situations where noise floor or interfering events exist. Second, we applied Squeeze-and-Excitation block on channel and frequency dimensions to efficiently extract feature characteristics. Result of our trained models on the STARSS22 test dataset achieved the best ER, F1, LE, and LR of 0.53, 49.8%, 16.0deg., and 56.2% respectively.
While many deep learning methods on other domains have been applied to sound event detection (SED), differences between original domains of the methods and SED have not been appropriately considered so far. As SED uses audio data with two dimensions (time and frequency) for input, thorough comprehension on these two dimensions is essential for application of methods from other domains on SED. Previous works proved that methods those address on frequency dimension are especially powerful in SED. By applying FilterAugment and frequency dynamic convolution those are frequency dependent methods proposed to enhance SED performance, our submitted models achieved best PSDS1 of 0.4704 and best PSDS2 of 0.8224.
A new database of head-related transfer functions (HRTFs) for accurate sound source localization is presented through precise measurement and post-processing in terms of improved frequency bandwidth and causality of head-related impulse responses (HRIRs) for accurate spectral cue (SC) and interaural time difference (ITD), respectively. The improvement effects of the proposed methods on binaural sound localization cues were investigated. To achieve sufficient frequency bandwidth with a single source, a one-way sealed speaker module was designed to obtain wide band frequency response based on electro-acoustics, whereas most existing HRTF databases rely on a two-way vented loudspeaker that has multiple sources. The origin transfer function at the head center was obtained by the proposed measurement scheme using a 0 degree on-axis microphone to ensure accurate spectral cue pattern of HRTFs, whereas in the previous measurements with a 90 degree off-axis microphone, the magnitude response of the origin transfer function fluctuated and decreased with increasing frequency, causing erroneous SCs of HRTFs. To prevent discontinuity of ITD due to non-causality of ipsilateral HRTFs, obtained HRIRs were circularly shifted by time delay considering the head radius of the measurement subject. Finally, various sound localization cues such as ITD, interaural level difference (ILD), SC, and horizontal plane directivity (HPD) were derived from the presented HRTFs, and improvements on binaural sound localization cues were examined. As a result, accurate SC patterns of HRTFs were confirmed through the proposed measurement scheme using the 0 degree on-axis microphone, and continuous ITD patterns were obtained due to the non-causality compensation. Source codes and presented HRTF database are available to relevant research groups at GitHub (https://github.com/han-saram/HRTF-HATS-KAIST).
2D convolution is widely used in sound event detection (SED) to recognize 2D patterns of sound events in time-frequency domain. However, 2D convolution enforces translation-invariance on sound events along both time and frequency axis while sound events exhibit frequency-dependent patterns. In order to improve physical inconsistency in 2D convolution on SED, we propose frequency dynamic convolution which applies kernel that adapts to frequency components of input. Frequency dynamic convolution outperforms the baseline model by 6.3% in DESED dataset in terms of polyphonic sound detection score (PSDS). It also significantly outperforms dynamic convolution and temporal dynamic convolution on SED. In addition, by comparing class-wise F1 scores of baseline model and frequency dynamic convolution, we showed that frequency dynamic convolution is especially more effective for detection of non-stationary sound events. From this result, we verified that frequency dynamic convolution is superior in recognizing frequency-dependent patterns as non-stationary sound events show more intricate time-frequency patterns.
Temporal dynamic models for text-independent speaker verification extract consistent speaker information regardless of phonemes by using temporal dynamic CNN (TDY-CNN) in which kernels adapt to each time bin. However, TDY-CNN shows limitations that the model is too large and does not guarantee the diversity of adaptive kernels. To address these limitations, we propose decomposed temporal dynamic CNN (DTDY-CNN) that makes adaptive kernel by combining static kernel and dynamic residual based on matrix decomposition. The baseline model using DTDY-CNN maintained speaker verification performance while reducing the number of model parameters by 35% compared to the model using TDY-CNN. In addition, detailed behaviors of temporal dynamic models on extraction of speaker information was explained using speaker activation maps (SAM) modified from gradient-weighted class activation mapping (Grad-CAM). In DTDY-CNN, the static kernel activates voiced features of utterances, and the dynamic residual activates unvoiced high-frequency features of phonemes. DTDY-CNN effectively extracts speaker information from not only formant frequencies and harmonics but also detailed unvoiced phonemes' information, thus explaining its outstanding performance on text-independent speaker verification.