![]() 5 show comparison of performance of different segmentation models with EfficientNetB3 backbone. Table 2: Network architectures used in this study 2.3 Experimentsįig. (e) Performance benchmark of encoder-backbones investigated in this study (MobileNetv2, EfficientNetB3, InceptionResNetv2) on ImageNet dataset tan2019efficientnet. Figure 2: Segmentation models investigated in the study: (a) UNet (b) Linknet (c) PSPNet (d) FPN Yakubovskiy:2019. We use the EfficientNetB3 architecture in this family which lies approximately in the middle of the spectrum as will be shown in Section 3.3. EfficientNet architectures use a compound scaling optimization with variable width, depth, resolution in order to optimize Accuracy and FLOPS. For analyzing trade-off between memory and accuracy over the complete design space, we have selected three backbones as representative workloads.InceptionResNetV2, a hybrid of two sophisticated networks, exhibits high accuracy but is computationally heavy, whereas MobileNetV2 despite having lower accuracy enables light-weight edge AI computing bianco2018benchmark. The prime motivation for the choice of backbones was to ensure an exhaustive exploration. Three encoder-backbones have been considered for this study: (i) EfficientNetB3 tan2019efficientnet, (ii) InceptionResnetV2 szegedy2017inception, (iii) MobileNetV2 sandler2018mobilenetv2. Table 1: Distribution of class samples in the dataset 1 shows an image chip and corresponding label map from the dataset. Maximum image size available in the dataset is 637 MB. ![]() A description of the class-wise distribution alongwith color map is provided in Table 1. For this study, we have only used raw RGB TIFFs, in order to demonstrate generalized capability without the need of additional channels such as elevation (as in the case of this dataset) or hyper-spectral bands (as in the case of other UAV datasets) due to relatively high costs of lightweight multispectral cameras yao2019unmanned. The label maps are annotated with 7 classes - namely Building, Clutter, Vegetation, Water, Ground, Car and ‘Ignore’ - the last class referring to missing pixels/ boundaries. Further they also generally have lower number of annotated classes inria.įor the purpose of this study, we have used DroneDeploy Dataset dronedeploy, comprising of 55 RGB images, along with single-channel elevation maps and label maps. ![]() While there exist many UAV video datasets Girisha_2019 avola2018uav uavid_isprs, UAV static images datasets are typically more application oriented Yang_2020, and hence suited for object detection applications stanforddronedataset.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |