3.2 Consideration
37 In relation to integer (1.5), the appellants submit that a "processing and control means arranged to receive image information from the light receiving means" should be understood to refer broadly to the processing and control means receiving any output from the light receiving means, such as the output from the sensors of a CCD or a camera. The term "recognise the presence of" means no more than the ability to identify the existence of a shadowed region which has been "cast by obstructions in the region". They submit that the word "obstructions" in this context means any object which blocks light from the light emitter, thereby creating a "shadowed region" on the light receiving means. Insofar as the primary judge took a contrary view, the appellants submit that he fell into error.
38 In our view, there is force in the criticisms of this aspect of the learned primary judge's reasoning.
39 His Honour accepted that "image information" is received from the light receiving means, such as a CCD or a projection screen, and considered that such information is used to "recognise the presence of one or more shadowed regions". This, he considered, does not mean any information whatsoever about an image, but that "the shape of the obstruction must be capable of being determined and sufficient to be used by the system" (at [229]). The primary judge said, that "recognise" in context (at [230]):
...means that the system is not just detecting a shadowed region, but the system recognises, that is, compares the shadowed region detected with what is previously known - the shape of the obstructions. This is consistent with the recitation in the Patent (at p 6 line 23 to p 7 line 14), particularly the comparison of the 'current image' with the 'stored image to determine the presence of any shadows' (at p 7 lines 9-11). The reference in integer 1.5 to the 'one or more shadowed regions' accords with the notion that the system can determine the shape of the obstruction.
40 We respectfully disagree with the learned primary judge's construction in this respect. No language in claim 1 indicates that the words "image information" and "recognise" are to be understood as inviting a comparison between the detected shadowed region and a previously known shape or obstruction. The words "image information", "recognise" and "obstruction" are not terms defined in the body of the specification so as to invite that conclusion. The words are of ordinary English. No party submitted that they involve terms of art that warrant, as aids to construction, evidence from a person skilled in the art. Accordingly, this Court on appeal is able to consider the question of construction.
41 The words "recognise the presence of one or more shadowed regions" indicate that "recognise" is to be understood as "to perceive as existing" rather than as inviting comparison with something previously known. It is simply the presence of the shadowed region that is required. This accords with the third definition of "recognise" in the Macquarie Dictionary, quoted by the primary judge at [200]. The first definition, which the primary judge preferred, is "to know again; perceive to be identical with something previously known". However, to adopt that meaning is to set to one side the words "the presence of" that immediately follow.
42 In our respectful view, the learned primary judge in this respect fell into error by adding a gloss to the language of the claim by incorporating the concept identified in the body of the specification at page 6 line 23 to page 7 line 14, which we have quoted in [15] above and to which his Honour referred. In this regard, the additional references by his Honour in [234] and [235] indicate an assumption that the scope of claim 1 is confined to the scope of the store and compare method and/or the shadow map method. The principles of patent and claim construction have been set out in many authorities, and it is not necessary to repeat them here; see Martin v Scribal Pty Ltd (1954) 92 CLR 17 at 59, Welch Perrin & Co Pty Ltd v Worrel (1961) 106 CLR 588 at 610, Populin v HB Nominees Pty Ltd (1982) 41 ALR 471 at 476, Fisher & Paykel Healthcare Pty Ltd v Avion Engineering Pty Ltd (1991) 103 ALR 239 at 255, Decor Corporation Pty Ltd v Dart Industries Inc (1988) 13 IPR 385 at 400. Although the claims are to be construed in the context of the specification as a whole, it is not legitimate to narrow or expand the boundaries of monopoly as fixed by the words of a claim by adding to those words glosses drawn from other parts of the specification; Welch Perrin at 610.
43 Whilst it is legitimate to refer to the rest of the specification to explain the background to the claims, to ascertain the meaning of technical terms and resolve ambiguities in the construction of the claims, it is not legitimate to narrow or expand the boundaries of monopoly as fixed by the words of a claim by adding to those words glosses taken from other parts of the specification; Welch Perrin at 610; Flexible Steel Lacing Co v Beltreco Ltd [2000] FCA 890 at [74], [75], and authorities there cited.
44 The importance of the actual language used in the claim to define the invention was emphasised by the Full Court in Fresenius Medical Care Australia Pty Limited v Gambro Pty Limited [2005] FCAFC 220 at [94]:
The whole of the specification must be read in order to construe the claim. So much is not in dispute. However, that does not mean that the words of the claim are to be ignored. Nor can it be ignored that the patentee has chosen not to include in a claim matters, integers or aspects of the invention that have been referred to in the body of the specification. This accords with the well-accepted principle enunciated in Welch Perrin at 610 that glosses drawn from the body of the specification cannot be used to narrow or expand the boundaries of the monopoly as fixed by the words of a claim.
45 Claim 1 makes no reference to a comparison of shapes or of shadow maps. The learned primary judge incorporated those concepts into the claim by having regard to the language used in relation to two of the embodiments described in the body of the specification, the store and compare method and the shadow map method. But, in our respectful view, it strains the language used by the patentee to import those concepts into claim 1. That claim requires that the processing and control means be arranged to receive "image information" from the light receiving means and "thereby recognise the presence of one or more shadowed regions". In context, as we have said, "recognise" does not equate to "compare" in the sense that one compares, as the primary judge says, the shadowed region detected with what is previously known - the shape of the obstructions.
46 Lazer Safe defends the construction adopted by the primary judge by contending that the passages identified in the specification are not simply an expression of a preferred embodiment, but must be read to be part of the statement of the invention, mandating the primary judge's construction of claim 1.
47 In our view, this submission is not supported by the language used in the specification. The Figures are introduced as describing the invention "by way of example", not definitively. The Description of the Invention introduces, by reference to the elements of Figure 4, "the embodiment shown". The scope of the invention the subject of the monopoly is determined by the claims. They delimit the boundary of the conception that represents the invention. In the present specification, the portion of the Patent that gives a description of the invention in the form of "the embodiment shown" is to be understood to be a description of a part or the whole of a product that could fall within the conception of the invention as claimed in one or more claims. The broadest aspect of the invention described is set out in the "Summary of the Invention", which, as we have noted, is set out in language corresponding with claim 1. Taken together, these matters direct the reader to an understanding that whilst the Description of the Invention provides the function of describing aspects of the invention and how they may be implemented, those aspects do not constrain the scope of the claims. Furthermore, as the bolded words within the passage identified at page 6 line 23 to page 7 line 14 set out above at [15] indicate, various alternatives are expressed as reflecting different ways in which the invention may be implemented. These provide a further indication that the body of the specification does not there purport to set out the scope of the invention.
48 Having regard to these considerations, in our view integer (1.5) does not require more than the safety system to include a processing and control means arranged to receive image information from the light receiving means and thereby detect the presence of the one or more shadowed regions on the light receiving means cast by obstructions in the illuminated region.
49 The appellants next challenge the primary judge's findings in relation to integer (1.6). In this regard, it is important to note that his Honour's reasoning involved a primary finding and a secondary or alternative finding, both of which he took into account when considering the question of infringement.
50 In his primary finding, his Honour concluded that the reference in integer (1.6) to "sufficient image information to determine the boundaries of the or each shadowed region and control movement of the part dependent on said image information" is to be understood to be consistent in a common sense way with the focus of the claim on "shape/outline recognition" (at [233]) and that the reference to "boundaries" reinforces "that the shapes or outlines of the obstructions must be determined sufficiently to be used by the system to control the tool" (at [234]). This was, in his Honour's view, confirmed by the passage at page 6 line 22 to page 7 line 14 and by the references to the processing and control means "creating 'a total picture made up of the image information' which is referred to as a 'shadow map' and which is compared to 'known safe shadow maps'" (at page 8 lines 1 - 12). These matters led the primary judge to the conclusion at [235] that:
Integer 1.5 and 1.6 of claim 1 require the safety system as claimed to engage in recognition of regions of obstructions by reference to shadow maps and their shapes, and require the comparison with either stored known safe or unstored unknown hazardous shadows of obstructions.
51 The secondary or alternative construction is that a boundary "encloses a region or area, extending around the entire perimeter of an object". Thus, his Honour found at [238] that even if his primary construction was wrong:
...the proper construction of claim 1 requires the boundary determination of an obstruction claimed by recognising a realistically substantial portion or part of the shadowed region, so as to determine a sizable portion of the outline or shape of the obstruction, in contrast to a dot, spot, point or edge of any such region.
52 The appellants contend that both of these constructions reflect error. We agree as to the primary construction, but not in relation to the secondary construction.
53 The appellants submit that integer (1.6) directs attention to the quality of the image information that is made available to the processing and control means and that the image information must be "sufficient" to determine the boundaries of the or each shadowed region. The integer is said to direct attention to the illumination of the region. This requires the image receiver being illuminated with light of sufficient intensity that a distinction can be made between the output from "light" pixels. The appellants contend that the determination of a boundary involves the identification of a transition from "light" to "dark" pixels (referred to as finding a "gradient"). Where a gradient is found between two adjacent pixels, an "edge" exists. The appellants submit that an edge is part of a boundary, but an edge alone is not sufficient to constitute a boundary within the proper construction of that term in claim 1. A boundary of a shadowed region is a line that demarcates the area of the shadow that is cast. As an area is two-dimensional, it must be possible to find gradients and determine the boundary in two dimensions. Therefore, the appellants submit, the phrase "sufficient image information to determine the boundaries of the or each shadowed region" in integer (1.6) requires that there be sufficient image information (that is, enough pixels at a suitable resolution and spacing) to allow the gradients (that is, the transition from light to dark pixels) to be found around the perimeter of the shadowed region. If the shadowed region comprises only one "dark" pixel, this requirement is met if there is sufficient image information to identify this transition from light to dark between one darkened pixel and the surrounding pixels.
54 In their written reply submissions, the appellants contend that the steps required by integer (1.6) are exemplified as follows:
(1) The output of the image receiver (such as a camera) being a two dimensional pixel matrix of adequate resolution for boundary determination;
(2) This output being provided from the light receiver to a processor;
(3) The image receiver being illuminated with light of sufficient intensity that a distinction can be made between the output from "light" pixels (i.e. not shadowed) and the output of "dark" pixels (i.e. shadowed); and
(4) The output from the light receiver being sufficient to be used in the control of the moving part (i.e. the tool of a press brake).
55 Lazer Safe submits that the requirement in integer (1.6) that there be "sufficient" image information to determine the boundaries of the or each shadowed region does not, as the appellants contend, set a standard purely on the quality of information. Rather, it involves the actual determination of the boundaries. It submits that the primary construction adopted by the Court below was consistent, as the learned primary judge said, in "a common sense way" with the focus of the claim on shape recognition. Further, Lazer Safe relies on the alternative or secondary construction adopted by the primary judge.
56 In our view, the material construction issue in relation to this aspect of integer (1.6) is: what is meant by the requirement that the illumination of the region is such that the processing and control means has sufficient image information to determine the boundaries of the or each shadowed region?
57 During the course of oral argument on the appeal, it became apparent that there was no real dispute between the parties as to the meaning of "boundaries". Mr Holler, who appeared for the appellants, accepted that in accordance with the dictionary definition, the "boundary" is something that serves to indicate the bounds or limits of something. Mr Fitzgerald QC, who appeared for Lazer Safe, relied on the evidence of Mr Acheson to the effect that the integer referred to a system capable of determining the perimeter of the shadow regions by finding the gradient in two dimensions between adjacent pixels or segments.
58 The primary construction adopted by the learned primary judge was influenced by his view that the words "sufficient image information" in integer (1.6) required enough information to determine the outline of the obstruction to enable it to be compared with either stored safe or unstored, unknown, hazardous shadows of obstructions; at [235]. However, that construction relies on the proposition that words "recognise the presence" in the context of integer (1.5) import the requirement of a comparison to be made between a stored known image and the shadowed region. We have rejected that approach for the reasons stated above. We do not consider that the language of integer (1.6) would justify the implication of such a comparison.
59 However, in our view the requirement of integer (1.6) that the illumination of the region be such that the processing and control means has sufficient information to determine the boundaries of the or each of the shadowed regions indicates that the processing means must have sufficient information to be capable of controlling the movement of the part dependent on identifying the bounds or limits of the shadowed region.
60 Accordingly, in our view there is force in the appellants' criticism of the first of the findings of the primary judge, but not the second. We have explained our view that integer (1.5) does not import with it a requirement that there be a comparison between the shape or a shadow map and the image of a shadowed region. The first construction of integer (1.6) advanced by the primary judge reiterates the requirement applied to integer (1.5) that there be a recognition of obstructions by reference to shadow maps and their shapes and requires a comparison with stored, known, safe or unstored, unknown, hazardous shadows of obstructions (at [235]). The requirement of a comparison is not present in integer (1.5) and cannot be discerned from the language used in integer (1.6). In our respectful view, the error that we have identified in relation to integer (1.5) permeates integer (1.6).
61 However, that error is not apparent in the second, alternative construction that the learned primary judge adopts for integer (1.6). The language of this integer requires that the illumination of the region be such that the processing and control means has sufficient information to first, determine the boundaries of the or each shadowed region, and secondly, control movement of the part dependent on the image information. The word "boundaries" has no antecedent in the language of the specification and is to be understood in the sense identified by the primary judge. Namely, as requiring identification of a realistically substantial portion or part of the shadowed region so as to determine the outline or shape of the obstruction.
62 Having regard to the construction that we consider to be correct, it is appropriate to turn to the 3 grounds of appeal that address this subject.
63 Ground 3 contends that the learned primary judge was in error insofar as he found that the invention claimed in the Patent concerned, and was limited to, a safety system involving the recognition of and comparison with stored images. Ground 4 is in materially the same form, in that it contends that the primary judge erred in construing claim 1, and in particular integers (1.5) and (1.6), as requiring comparison between regions of obstructions by reference to stored known safe obstructions. For the reasons explained, we find that these grounds have been established.
64 However, in ground 5 the appellants contend that the learned primary judge erred by requiring the boundary determination of an obstruction to involve the recognition of a "realistically substantial portion or part of the shadowed region" so as to determine a sizeable portion to the outline or shape of the obstruction "in contrast to a dot, spot, point or edge of any such region". In our view the appellants have not demonstrated error on the part of the primary judge in this respect. This was the alternative or secondary construction adopted by the primary judge in construing integer (1.6).
65 Finally, in ground 6 the appellants contend that in construing claim 1, the primary judge erred at [239] by failing to construe claim 1 in the context of the dependent claims, and in particular in the context of claim 2. Having regard to the conclusion that we have reached in relation to grounds 3, 4 and 5, it is unnecessary to address this ground in detail. The appellants submit that the primary judge's construction of integers (1.5) and (1.6) is not "harmonious" with the scope of dependent claims 2 and 8 - 11. The appellants do not, however, submit that the construction adopted by the primary judge renders those claims redundant. If that were the case, then the Court would endeavour to construe claim 1 in a manner that avoided redundancy; see Bodkin, Patent Law in Australia, 3rd ed, 2019, at [21910]. This is consistent with the proper approach to claim construction that construes claims in the context of the specification as a whole, including the other claims. Nevertheless, if the plain meaning of a claim has the effect that one or more dependent claims are redundant, then the consequence of redundancy would not drive the construction of the claim itself. In Nichia Corporation v Arrow Electronics Australia Pty Ltd [2019] FCAFC 2 (Besanko, Jagot and Nicholas JJ), Jagot J noted (Besanko and Nicholas JJ agreeing) that there was a degree of overlap (or redundancy) between two claims, but declined to use this as a basis to prefer a particular construction. Her Honour held at [41]:
Nor do I consider that the existence of a degree of overlap between claims 3 and 6 provides a sound basis to prefer one construction of claims over another. Even if, as the primary judge said at [97], claim 6 is "largely redundant" because claim 3 also covers two or more fluorescent materials of different compositions represented by the general formula, the redundancy is not complete as claim 6 specifically defines and claims a sub-set of claim 3 devices. In any event, some, even extensive, redundancy in the drafting of patent claims seems more an indicator of an abundance of caution than a guide to the meaning of "contains" which appears in both claims 3 and 6 (and other claims) and thus should generally be construed as having the same meaning wherever it is used.
66 It may be that the particular language of an earlier claim has the consequence that later claims were poorly drafted and are accordingly redundant.
67 It is in this legal context that the appellant's submission that the construction adopted by the learned primary judge is not "harmonious" with the dependent claims must be viewed. The appellants do not point to any particular discord (to extend the metaphor) between the construction and the dependent claims that renders the construction adopted by the primary judge incorrect. For the reasons given, we consider that the alternative construction adopted by the primary judge of integer (1.6) reflects a proper understanding of the meaning of the language of claim 1, read in the context of the entirety of the specification including the dependent claims. Furthermore, we consider that it does not place a gloss on the language of that claim by placing undue emphasis on the language of the specification, and in particular by selecting one or other of the preferred embodiments.
68 We now turn to address the question of infringement, having regard to the construction of claim 1 that we have found to be correct.