Method and apparatus for selecting image creatives

Abstract

The embodiment of the invention provides a method and device for selecting a picture advertisement material, which is conducive to overcoming various deficiencies in picture advertisement material selection in the prior art. The method comprises the steps of storing multiple items of training data, wherein each item of training data comprises a picture advertisement material of an advertisement and a score of the advertisement when being accessed at each time, and whether the score is positive or negative is determined according to whether the advertisement is converted or not after being currently accessed; training the training data by using a convolutional neural network to obtain a model, wherein an input layer of the convolutional neural network is acquired according to a pixel value of the picture advertisement material, a convolution kernel in a convolution layer is preset, and the weight of each edge in the convolutional neural network and the convolution kernel are updated according to the difference between a score outputted by the model and the score in training; and calculating the score of a specified picture advertisement material by using the model, and determining whether to select the specified picture advertisement material or not according to the score.

Description

technical field [0001] The invention relates to the technical field of computers and software, in particular to a method and device for selecting picture advertisement materials. Background technique [0002] The material (usually a picture) is the basic raw material of the advertising system, and the quality of the material directly determines the comprehensive income of the advertising platform and the advertiser (clicks, conversions, etc. brought by the advertisement). Therefore, the reasonable selection of advertising materials for advertisers is a very critical part of the advertising system, and they hope to optimize the selection of materials for image ads. Therefore, the materials in the following are all image types. [0003] Advertising platforms in the industry such as Facebook, Google, Baidu, and Tencent will provide some suggestions for image creative optimization, such as: material clarity, size, template reference, etc. [0004] At present, the common way for...

AI Technical Summary

Problems solved by technology

[0013] From the advertiser's point of view: (1). The cycle of material selection is long, relying on the cumulative amount of material played by the advertising playback system, and the quality of the material cannot be accurately measured when the amount of playback is insufficient; (2). Online delivery test for some low-quality materials , will result in a loss of profit

[0014] From the perspective of the advertising system: (1). Advertisers putting in a large amount of low-quality materials will increase the advertising retrieval burden of the advertising playback system; (2). Frequent uploading and offline of materials will increase the instability of the advertising system playback effect.

Images