Image recognizing method for preventing recognition results from confusion

[0052]FIG. 6A is a diagram of the first embodiment showing confusion of an object recognition result. The parent-categories shown in FIG. 6A, such as a Phone category, a Tablet category, a Laptop category and a Monitor category all have a common feature that is the Monitor feature, so their recognition results may cause confusion. As a result, if the multiple targets to be recognized and inputted by the user include these parent-categories simultaneously, the recognition platform may use the child-classifiers corresponding to a Phone monitor subcategory, a Tablet monitor subcategory, a TV monitor subcategory, a Laptop subcategory, etc., to perform the analysis action to the video, so as to prevent the recognition results from mis-recognizing phones, tablets, TVs or laptops as a monitor.

[0053]FIG. 6B is a diagram of the second embodiment showing confusion of an object recognition result. The parent-categories shown in FIG. 6B, such as a Laptop category, a PC category and a Keyboard category all have a common feature that is the Keyboard feature, so their recognition results may cause confusion. As a result, if the multiple targets to be recognized and inputted by the user include these parent-categories simultaneously, the recognition platform may use the child-classifiers corresponding to a Laptop keyboard subcategory, a PC keyboard subcategory, etc., to perform the analysis action to the video, so as to prevent the recognition results from mis-recognizing laptops or PCs as a keyboard.

[0054]FIG. 6C is a diagram of the third embodiment showing confusion of an object recognition result. The parent-categories shown in FIG. 6C, such as an Automobile category, a Bicycle category and a Wheel category all have a common feature that is the Wheel feature, so their recognition results may cause confusion. As a result, if the multiple targets to be recognized inputted by the user include these parent-categories simultaneously, the recognition platform may use the child-classifiers corresponding to an Automobile wheel subcategory, a Bicycle wheel subcategory, etc., to perform the analysis action to the video, so as to prevent the recognition results from mis-recognizing automobiles or bicycles as a wheel.

[0055]FIG. 6D is a diagram of the first embodiment showing confusion of a scene recognition result. The parent-categories shown in FIG. 6D, such as a Restaurant category, a BAR category and a Decoration category all have a common feature that is the Decoration feature, so their