Method and device for extracting information from pie chart

An electronic device for extracting information from a pie chart for display, comprising:

• a processor;
• memory;
• and
• a plurality of computer instructions stored in the memory, wherein the computer instructions, when executed by the processor, cause the electronic device to perform operations including:
• detecting, by the electronic device, each element in a pie chart to be processed and position information thereof, the elements comprising text elements and legend elements;
• performing, by the electronic device, text recognition on the detected text elements and legend elements to obtain text information corresponding to the text elements and legend texts included in the legend elements respectively;
• and
• obtaining, by the electronic device, sector information and legend information according to each detected element and position information thereof and the legend texts, and enabling, by the electronic device, the sector information to correspond to the legend information one by one, wherein the sector information comprises a sector color and a proportion of the sector in the pie chart, and the legend information comprises a legend color and a corresponding legend text thereof,
• wherein each element in the pie chart to be processed is detected by the electronic device by adopting a target detection method of a Faster R-CNN model and the Faster R-CNN model is pre-stored in the electronic device and derived via the following training method:
• step 1. randomly initializing learning parameters in the Faster R-CNN model;
• step 2. inputting a batch of training samples to the Faster R-CNN model to obtain a predicted classification probability pi and a predicted bounding box coordinate vector ti of the ith element box in the training samples under the current model parameters;
• step 3. performing loss calculation on the output result in step 2 by adopting the following formula, and solving the average loss L of all the element boxes in the batch of training samples,
• L
• ⁡
• (
• {
• p
• i
• }
• ,
• {
• t
• i
• }
• )
• =
• 1
• N
• cls
• ⁢
• ∑
• i
• ⁢
• ⁢
• L
• cls
• ⁡
• (
• p
• i
• ,
• p
• i
• *
• )
• +
• λ
• ⁢
• ⁢
• 1
• N
• reg
• ⁢
• ∑
• i
• ⁢
• ⁢
• p
• i
• *
• ⁢
• L
• reg
• ⁡
• (
• t
• i
• ,
• t
• i
• *
• )
• ,
• in which L({pi},{ti}) is the loss of the ith element box, Ncls is the value of mini-batch, Nreg is the number of anchor positions, λ is weight,
• ∑
• i
• ⁢
• ⁢
• L
• cls
• ⁡
• (
• p
• i
• ,
• p
• i
• *
• )
• is the logarithmic loss of a target and a non-target, and
• ∑
• i
• ⁢
• ⁢
• p
• i
• *
• ⁢
• L
• reg
• ⁡
• (
• t
• i
• ,
• t
• i
• *
• )
• is a regression loss function;
• step 4. solving the minimum L, and updating all the learning parameters in the Faster R-CNN model;
• and
• step 5. repeatedly executing steps 2 to 4 until reaching a set number of iterations.

A non-transitory computer readable storage medium comprising computer readable instructions that, when executed by a processor of an electronic device having a display, cause the electronic device to perform operations including:

• detecting, by the electronic device, each element in a pie chart to be processed and position information thereof, the elements comprising text elements and legend elements;
• performing, by the electronic device, text recognition on the detected text elements and legend elements to obtain text information corresponding to the text elements and legend texts included in the legend elements respectively;
• and
• obtaining, by the electronic device, sector information and legend information according to each detected element and position information thereof and the legend texts, and enabling, by the electronic device, the sector information to correspond to the legend information one by one, wherein the sector information comprises a sector color and a proportion of the sector in the pie chart, and the legend information comprises a legend color and a corresponding legend text thereof,
• wherein each element in the pie chart to be processed is detected by the electronic device by adopting a target detection method of a Faster R-CNN model and the Faster R-CNN model is pre-stored in the electronic device and derived via the following training method:
• step 1. randomly initializing learning parameters in the Faster R-CNN model;
• step 2. inputting a batch of training samples to the Faster R-CNN model to obtain a predicted classification probability pi and a predicted bounding box coordinate vector ti of the ith element box in the training samples under the current model parameters;
• step 3. performing loss calculation on the output result in step 2 by adopting the following formula, and solving the average loss L of all the element boxes in the batch of training samples,
• L
• ⁡
• (
• {
• p
• i
• }
• ,
• {
• t
• i
• }
• )
• =
• 1
• N
• cls
• ⁢
• ∑
• i
• ⁢
• ⁢
• L
• cls
• ⁡
• (
• p
• i
• ,
• p
• i
• *
• )
• +
• λ
• ⁢
• ⁢
• 1
• N
• reg
• ⁢
• ∑
• i
• ⁢
• ⁢
• p
• i
• *
• ⁢
• L
• reg
• ⁡
• (
• t
• i
• ,
• t
• i
• *
• )
• ,
• in which L({pi},{ti}) is the loss of the ith element box, Ncls is the value of mini-batch, Nreg is the number of anchor positions, λ is weight,
• ∑
• i
• ⁢
• ⁢
• L
• cls
• ⁡
• (
• p
• i
• ,
• p
• i
• *
• )
• is the logarithmic loss of a target and a non-target, and
• ∑
• i
• ⁢
• ⁢
• p
• i
• *
• ⁢
• L
• reg
• ⁡
• (
• t
• i
• ,
• t
• i
• *
• )
• is a regression loss function;
• step 4. solving the minimum L, and updating all the learning parameters in the Faster R-CNN model;
• and
• step 5. repeatedly executing steps 2 to 4 until reaching a set number of iterations.