Fujifilm vs Motorola – was this predictable from Cluster Searching?

Fujifilm vs Motorola (California Northern District Court, Case No. 3:12-cv-03587) wrapped up last week, with the Jury ordering Motorola to pay Fujifilm US$10.2 million for its infringement of claims 1, 7 and 11 of US6144763. The jury also found infringement of US6915119, but then that the asserted claims 1, 13, and 35 of this patent were obvious. 

Altogether, five patents were asserted by Fujifilm in this case:

  • US6144763 Method and apparatus for compression coding of image data representative of a color image and digital camera including the same, filed 1998 ( found to be valid and infringed by the jury) 
  • US7327886, Photographing apparatus, method and program, filed 2005, considered by jury but found to be invalid and not infringed. This covers the abilty to recognise if a face is found in an image, and is useful as a prerequsite for red-ey correction.
  • US6915119  Telephone and data transmitting method for telephone, filed  2000, found to be infringed but invalid
  • US5734427 High resolution electronic still camera with an electronic viewfinder for displaying a reduced image, filed 1995, listed in initial complaint but not considered by the Jury
  • US8306285 Photographing apparatus, method and program, filed in 2011, considered by the jury but found to be invalid and not infringed. This covers the abilty to recognise if a face is found in an image. 

But how much of this could have been predicted by patent analysis? Ambercite has just released Cluster Searching, which can be very helpful for this sort of licensing analysis. In this case, it is straightforward to run a query for the most predicted similar ~2000 patents to this starting set of patents. This query is very easy to set up and simply requires the patent numbers and date cut-off. In this case we might use the March 1994 priority date of the oldest of the Fujifilm patents.

StandardQuery.jpg

 

The magnifying glass starts the search process, and the query takes a few seconds to run. The result is the following list of similar patents, ranked in order of their similarity. Some of the fields are obvious, but others need an explanation.

Output_20150511-065647_1.jpg

These fields are:

(1) Rank – the overall ranking of the patents we found

(2) Similarity – our metric predicting the similarity of the patents we found to the starting (‘seed’) patents. The ranking is based on these values, and if two patents have the same similarit value, also on the ‘AmberScore’ value (see explanation below)

(3) L.P., short for “Licensing Potential“. This combines the Similarity value and AmberScore value to suggest which patents or owners commercialisation managers should look at first

(4) The blue and orange link buttons open up further patent details in either Google Patent or Espacenet (blue button) or show the network of each patent in AmberScope (orange button)

(5) Citation, shows if the found patent is a direct citation link of one of the seed patents, or only linked indirectly  (i.e. connected to the seed patents via another patent.

(6) AmberScore – this is our unique metric that predicts the commercial potential of a patent based on its citation connections. The average granted US patent has an AmberScore value of 1.0 and the higher the value the better.

Reading these results, the most similar patent to the five seed patents is US7991200, for a Method and apparatus for compression coding of image data representative of a color image and digital camera including the same, also filed by Fujifilm, The next most similar patent was filed by Toshiba, covering a Person recognition apparatus.

 

What about Licensing Potential?

From a licensing managers’ viewpoint, though, it is perhaps more interesting to rank the results by “Licensing Potential”. Each column can be sorted by the results by clicking on the column heading. If  we sort for the patents with the highest licensing potential, we see the following results

 

Results-sorted-by-LP_20150511-065850_1.jpg

This suggests that the patent with the highest licening potential is US5687169, filed in 1995 for a Full duplex ultrawide-band communication system and method, and now owned by Alereon Limited. A key reason for this ranking is its very high AmberScore value of 46.77, an excellent AmberScore value. This is an important patent.

However ideally, such analysis would be based on all1733 patents we found, and not just the top 10 patents. This is easy to do if the results are exported to Excel using the “Export to Excel” button we provide. 

Once exported to Excel, it is then straightforward to work outn which companies have the greatest total Licensing Potential, as is shown in the table below.

Rank

Patent owners

Total Licensing Potential

Number of patents in list

1

Fotonation

224.5

107

2

Eastman Kodak

85.7

43

3

Microsoft

79.1

65

4

Canon

77.1

100

5

Fujifilm

74.1

74

6

Sony

45.9

64

7

Ricoh

42.9

38

8

HP

42.1

29

9

Nokia

33.6

21

10

AT & T

33.2

27

11

Toshiba

28.7

21

12

Xerox

28.5

22

13

Sharp

28.1

17

14

Panasonic

26.6

35

15

Ericsson

25.3

15

16

Seiko Epson

23.4

22

17

Alereon, inc.

22.6

4

18

NEC

22.4

25

19

Philips

21.3

18

20

Intel

19.8

17

21

Samsung

17.6

23

22

IBM

17.0

16

23

Qualcomm

15.8

22

24

Motorola

12.8

15

25

Kyocera

12.8

7

26

Siemens

12.3

9

27

Sanyo

11.8

15

28

Texas Instruments

11.6

8

29

Nikon

11.3

13

30

Konica Minolta

11.0

15

 

There are two ways we could look at this list. In the first way, we just focus on the top ranked companies, ie. Fotonation, the now defunct Eastman Kodak, Microsoft etc. 

A second way of looking at this to make the obvious observation that a lot of the most interesting things going on in the world of imaging is happening with smartphones. And indeed in this list there a number of companies that make smartphones, i.e Microsoft, Sony, Nokia, Samsung and Motorola – along with several companies that make components for smartphones. 

But this is only a start of the process. For example, for each of the companies on this list, we could look at some of the highest ranked (by Licensing Potential or by similarity, or both) patents to see what these patents could tell us about the activites of these companies. For example

  • Fotonation: One of the highest ranked patents by Licensing Potential is US7916897, for Face tracking for controlling imaging parameters  (Licensing potential of 6.0), while the most similar patent is US640777 for a Red-eye filter method and apparatus
  • Microsoft: The highest ranked patent by Licensing Potential (and the most similar) is US6895112, for Red-eye detection based on red region detection with eye confirmation (Licensing potential of 4.1).
  • Motorola: The highest ranked patent by Licensing Potential is US5987325, for a Multiple smart card phone and method (Licensing potential of 3.1), while the most similar patent is US7653413 for Subscriber device and method therein for enhancing interfaces thereto

To be honest, neither of the Motorola patents mentioned are that similar to the asserted Fujifilm patents. It is a given that patents do not infringe patents, but patents filed by allegedly infringing patents can provide  guidance on what the alleged infringer may have been developing. In this case, there does not appear to be close similarity between these Motorola patents and the Fujifilm patents. This is not to say that products produced by Motorola were not infringing patents filed by Fujifilm, but this is consisten with the view of the jury only finding two of the asserted patents to be infringed. 

Equally, while a $10.2 million judgement is a lot of money by my personal standards, this equates to a one-off lump sum payment of 50 cents for each of the 20.4 million found-to-be-infringing devices sold by Motorola prior to June 2014. $10.2 million is not a material sum for either Motorola or Fujifilm –  and in the case of Fujifilm may be significantly reduced by its legal fees.

 

Summary

A review of five patents patent asserted by Fujifilm against Motorola using our advanced citation analysis found only a weak relationship between the most similar Motorola patents and the asserted patents. While this is not enough in itself to be a defence of Motorola, it is interesting that the jury for this case only found infringement by two of the five Fujifilm patents – one of which was invalidated. In the end only a low damages award was ordered by the jury.

 

Interested in trying Cluster Searching for yourself?

Ambercite is offering free two week trials for interesting parties – please contact us for details. Cluster Searching can be learnt in minutes, and we provide free support. 

Other applications for Cluster Searching include patent invalidation, freedom to operate, patentability and landscaping.

 

 

 

 

 

 

 

 

 

 

Read More