3 results about "Layers" patented technology

The invention relates to an image classification method based on a multi-core dense connection network. The method comprises the following steps: S1, establishing an image set; S2, constructing a multi-core density connection network model, wherein the multi-core density connection network model comprises an intensive connection unit, an attention unit and a classification unit, the intensive connection unit comprises at least two intensive connection modules, each intensive connection module comprises a plurality of bottleneck layers, each bottleneck layer comprises two convolution layers which are arranged in sequence, and each convolution layer comprises at least two convolution layers which are arranged in sequence, wherein the convolution kernels of the second convolution layer in thebottleneck layers in different dense connection modules are different in size; S3, training the multi-core dense connection network model to obtain a trained multi-core dense connection network model; and S4, inputting the test set into the trained multi-core dense connection network model, and outputting an image classification result. Compared with the prior art, the method has the advantages that depth features of different scales on the extreme image can be effectively extracted through convolution kernels of different sizes, and a better classification effect is achieved.

The invention discloses an optical color-changing anti-counterfeiting stay wire which comprises a base material, the surface of the base material is plated with multiple layers of transparent films ina vacuum mode, the transparent films are made of a color-changing material with a micro-prism effect, an image-text layer is tightly attached to the bottom surface of the base material, a laser layeris tightly attached to the bottom surface of the image-text layer, and a substrate is tightly attached to the bottom surface of the laser layer. A pressure-sensitive adhesive fills the joint of the laser layer and the substrate, the substrate is made of a transparent film material special for laser press polishing, a corona layer is formed on the surface of the substrate through corona treatment,the surface of the corona layer is coated with polyol resin, an image-text surface is arranged at the joint of the substrate and the image-text layer, and the image-text surface is located on the surface of the image-text layer. The image-text surface is a two-dimensional code surface layer formed by a multi-color film through multi-layer vacuum coating so as to facilitate counterfeiting prevention, and a laser surface is arranged at the joint between the image-text layer and the laser layer and located on the surface of the laser layer so as to enable the laser surface to be pressed and fixed on the bottom surface of the image-text layer.

The embodiment of the invention discloses an image filling method, device and equipment and a storage medium, and relates to the technical field of artificial intelligence, deep learning, computer vision and image filling, and the image filling method comprises the steps: carrying out the multi-layer coding of the features of a to-be-filled image, generating a multi-layer coding feature layer, andenabling the sizes of the multi-layer coding feature layer to be reduced layer by layer; enabling the multiple coding feature layers to be subjected to layer-by-layer decoding processing to obtain multiple decoding feature layers and a first image, wherein no missing area exists in the first image, and the layer-by-layer decoding processing comprises splicing operation on the decoding feature layers and the coding feature layers which are the same in size; and performing up-sampling processing on the first image to obtain a plurality of up-sampling feature layers and a second image subjectedto up-sampling processing optimization, the up-sampling processing including splicing operation on the up-sampling feature layers and the decoding feature layers of the same size. According to the invention, high-quality image filling can be realized.